Skip to main content

Full text of "Protocols_molecular_biology"

See other formats






Edited by 



Miep H. Helfrich, PhD 

ilston, MD 




^Humana Press 

I ■ I I I 


Human Osteoblast Culture 

James A. Gallagher 

1 . Introduction 

Osteoblasts are the cells responsible for the formation of bone; they 
synthesize almost all of the constituents of the bone matrix and direct its 
subsequent mineralization. Once a phase of active bone formation is com- 
pleted the osteoblasts do not become senescent but instead redifferentiate 
into one of two other cell types: osteocytes and bone lining cells, both of 
which play a major role in the regulation of calcium homeostasis and bone 

Researchers have endeavored to culture osteoblasts from human bone for 
several reasons: 

1 . To investigate the biochemistry and physiology of bone formation. 

2. To investigate the molecular and cellular basis of human bone disease. 

3. To investigate the role of cells of the osteoblastic lineage in regulating bone 

4. To screen for potential therapeutic agents. 

5. To develop and test new biomaterials. 

6. To use cell therapy in tissue engineering and bone transplantation. 

The structure of bone tissue, the heterogeneity of cell types, the cross- 
linked extracellular matrix, and the mineral phase combine to make bone 
a difficult tissue from which to extract cells. Consequently, early attempts 
to culture osteoblasts avoided human tissue and instead relied on enzy- 
matic digestion of poorly mineralized fetal or neonatal tissue from 
experimental animals. The first attempt to isolate cells from adult human 
bone, using demineralization and collagenase digestion, was reported by 
Bard and co-workers (I). The cultured cells were low in alkaline phos- 
phatase and collagen synthesis, which were then regarded as the best 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 

01/Gall/1-18/F1 3 1 2/26/03, 10:44 AM 

I ■ I I I 


4 Gallagher 

markers of the osteoblastic phenotype. Although the cells remained viable 
for up to 2 wk they did not proliferate, and it was concluded that osteocytes 
were the predominant cell type present. Mills et al. used the alternative 
approach of explant culture and were successful in culturing cell populations 
that included parathyroid hormone (PTH) responsive and alkaline phosphatase 
positive cells (2). 

The first successful attempts to isolate large numbers of cells that expressed 
an osteoblastic phenotype from human bone were undertaken in Graham 
Russell's laboratory at the University of Sheffield in the early 1980s. The 
defining characteristics of these studies were (1) the use of explant cultures, 
which avoided the need for digestion of the tissue and (2) the availability of an 
appropriate phenotypic marker. Successful culture of any cell type can be 
achieved only if there is a specific marker of the phenotype that can be used to 
confirm the identity of the cells in vitro. In this case, the marker was the then 
recently discovered bone gla protein as measured by a radioimmunoassay 
developed by Jim Poser (3,4). Nearly 20 yr later, bone gla protein, now known 
as osteocalcin, undoubtedly remains the most specific marker of the osteoblas- 
tic phenotype. 

Although this culture system has been extensively modified by several 
groups of researchers (see Note 1), the vast majority of published reports on 
isolation of human osteoblasts still essentially use this simple but highly repro- 
ducible explant technique. This technique and its modifications have been 
described, compared, and reviewed elsewhere (5,6). The aim of this chapter is 
to describe the basic methodology that is used in the author's laboratory. This 
is shown schematically in Fig 1. The nomenclature used by various research 
groups to describe the isolated cells includes "human bone cells," "human 
osteoblasts in vitro," "human osteoblastic cells," and "HOBS." We have pre- 
ferred the conservative term "human bone derived cells" (HBDCs), and this is 
used throughout this chapter. 

HBDCs have been widely used to investigate the biology of the human 
osteoblast, and their use has facilitated several major developments in our 
understanding of the hormonal regulation of human bone remodeling. These 
cells have also been used to investigate the cellular and molecular pathology of 
bone disease. The major milestones in the culture of human osteoblasts are 
summarized in Table 1. Figure 2 shows the increase in the application of 
human osteoblast cultures since the initial reports in 1984. 

Human bone cell culture is now becoming an important tool in tissue engi- 
neering to test the biocompatibility and osteogenicity of novel biomaterials 
and also for autologous transplantation of osteoblastic populations expanded 
in vitro. 

01/Gall/1-18/F1 4 I 2/26/03, 10:44 AM 

I ■ I I I 

Human Osteoblast Culture 


Dice bone 

cancellous bone 
with bone rongeurs 


(Passage 2) 


Remove explants 

dice with scalpeE 
and replate 

4-6 weeks replated 
explants give rise to 
confluent cultures 


Phenoiypic characterization 
(e.g., osteocalcin expression) 



Fig. 1. Technique used to isolate cells expressing osteoblastic characteristics 
(HBDCs) from explanted cancellous bone. El, explant 1; E2, explant 2. 

2. Materials 

2. 1. Tissue-Culture Media and Supplements 

1. Phosphate-buffered saline (PBS) without calcium and magnesium, pH 7.4 

2. Dulbecco's modification of minimum essential medium (DMEM) (Invitrogen) 
supplemented to a final concentration of 10% with fetal calf serum (FCS), 2 mM 
L-glutamine, 50 U/mL of penicillin, 50 u-g/mL streptomycin. Freshly prepared 50 
Hg/mL of L-ascorbic acid should be added to cultures in which matrix synthesis 
or mineralization is being investigated (see Note 2). 

3. Serum-free DMEM (SFM). 

4. FCS (see Note 3). 

5. Tissue culture flasks (75 cm 2 ) or Petri dishes (100-mm diameter) (see Note 4). 

2.2. Preparation of Explants 

1. Bone rongeurs from any surgical instrument supplier. 

2. Solid stainless steel scalpels with integral handles (BDH Merck). 


2/26/03, 10:44 AM 

I ■ I I I 



Table 1 

Phenotypic Milestones in the Culture and Characterization 

of Osteoblastic Cells (HBCDs) from Human Bone 

Isolation of viable cells from human bone 
Introduction of explant culture 
Production of osteocalcin 

High alkaline phosphatase activity 

Responsiveness to PTH 

Synthesis of type I but not type III collagen 
Synthesis of other bone matrix proteins 

Response to cytokines 

Response to oestrogen 

Expression of purinoceptors 

Production of nitric oxide 
Investigation of specific pathologies 

Formation of mineralised nodules 
Formation of bone in vitro and in vivo 

Bardetal., 1972(2) 
Mills et al., 1979 (2) 
Gallagher et al., 1984 (3) 
Beresford et al., 1984a (4) 
Gallagher et al., 1984 (3) 
Beresford et al., 1984 (4) 
Gehron-Robey and Termine 1985 (7) 
Auf'molk et al., 1985(8) 
Beresford et al., 1984 (4) 
MacDonald et al., 1984 (9) 
Gehron-Robey and Termine 1985 (7) 
Auf'mkolk et al., 1985 (8) 
MacDonald et al., 1986 (10) 
Beresford et al., 1986 (77) 
Gehron-Robey and Termine 1985 (7) 
Fedarko et al., 1992 (12) 
Beresford et al., 1984 (13) 
Gowenetal., 1985(74) 
Vaishnav et al., 1984 (15) 
Eriksenetal 1988(76) 
Schoefl et al., 1992(77) 
Bowler et al., 1995 (18) 
Ralston et al., 1994(79) 
Marie et al., 1988 (20) 
Walsh etal., 1995(27) 
Beresford et al., 1993 (22) 
Gundle et al., 1995 (23) 

2.3. Passaging and Secondary Culture 

1. Trypsin-EDTA solution: 0.05% Trypsin and 0.02% EDTA in Ca 2+ - and Mg 2+ - 
free Hanks' balanced salt solution, pH 7.4 (Invitrogen). 

2. 0.4% Trypan blue in 0.85% NaCl (Sigma Aldrich). 

3. 70-jxm "Cell Strainer" (Becton Dickinson). 

4. Neubauer hemocytometer (BDH Merck). 

5. Collagenase (Sigma type VII from Clostridium histolyticum). 

6. DNase I (Sigma Aldrich). 


2/26/03, 10:44 AM 

I ■ I I I 

Human Osteoblast Culture 


1980 1985 1990 1995 




Fig. 2. Graph showing the increase in the application of human osteoblast cultures 
since the initial reports in 1984. 

2.4. Phenotypic Characterization 

1. 1,25-Dihydroxy vitamin D 3 [l,25-(OH) 2 D 3 ] (Leo Pharmaceuticals or Sigma 

2. Menadione (vitamin K3) (Sigma Aldrich). 

3. Alkaline phosphatase assay kit (Sigma Aldrich). 

4. Staining Kit 86-R for alkaline phosphatase (Sigma Aldrich). 

5. Osteocalcin radioimmunassay (IDS Ltd., Boldon, UK) (see Note 5). 

6. Polymerase chain reaction (PCR) primers and reagents for a panel of osteoblastic 
markers including osteocalcin (IDS Ltd., Boldon, UK). 

2.5. In Vitro Mineralization 

1. Dexamethasone (Sigma Aldrich). 

2. Hematoxylin (BDH Merck). 

3. L-Ascorbic acid (see Note 2). 

4. Inorganic phosphate solution: Mix 500 mM solutions of Na 2 HP0 4 and NaH 2 P0 4 
in a 4:1 (v/v) ratio. Sterile filter and store at 4°C prior to use. 

2.6. Cryopreservation of Cells 

1. Dimethyl sulfoxide (DMSO) (Sigma Aldrich). 

2. Cryoampules. 

3. Cell freezing container. 


2/26/03, 10:44 AM 

I ■ I I I 


8 Gallagher 

3. Methods 

3. 1. Establishing Primary Explant Cultures 

A scheme outlining the culture technique is shown in Fig. 1. 

1. Transfer tissue, removed at surgery or biopsy, into a sterile container with PBS or 
serum-free medium (SFM) for transport to the laboratory with minimal delay, 
preferably on the same day (see Note 6). An excellent source is the upper femur 
of patients undergoing total hip replacement surgery for osteoarthritis. Cancel- 
lous bone that would otherwise be discarded is removed from this site prior to the 
insertion of the femoral prosthesis. The tissue obtained is remote from the hip 
joint itself, and thus from the site of pathology, and is free of contaminating soft 
tissue (see Note 7). 

2. Remove soft connective tissue from the outer surfaces of the bone by scraping 
with a sterile scalpel blade. 

3 Rinse the tissue in sterile PBS and transfer to a sterile Petri dish containing a 
small volume of PBS (5-20 mL, depending on the size of the specimen). If the 
bone sample is a femoral head, remove cancellous bone directly from the open 
end using sterile bone rongeurs or a solid stainless steel blade with integral 
handle. Disposable scalpel blades may shatter during this process. With some 
bone samples (e.g., rib), it may be necessary to gain access to the cancellous bone 
by breaking through the cortex with the aid of the sterile surgical bone rongeurs. 

4. Transfer the cancellous bone fragments to a clean Petri dish containing 2-3 mL 
of PBS and dice into pieces 3-5 mm in diameter. This can be achieved in two 
stages using a scalpel blade first, and then fine scissors. 

5. Decant the PBS and transfer the bone chips to a sterile 30-mL "universal con- 
tainer" with 15-20 mL of PBS. 

6. Vortex-mix the tube vigorously three times for 10 sec and then leave to stand for 
30 sec to allow the bone fragments to settle. Carefully decant off the supernatant 
containing hematopoietic tissue and dislodged cells, add an additional 15-20 mL 
of PBS, and vortex-mix the bone fragments as before. Repeat this process a mini- 
mum of three times, or until no remaining hematopoietic marrow is visible and 
the bone fragments have assumed a white, ivory-like appearance. 

7. Culture the washed bone fragments as explants at a density of 0.2-0.6 g of tissue/ 
100-mm diameter Petri dish or 75-cm 2 flask (see Note 4) in 10 mL of medium at 
37° in a humidified atmosphere of 95% air, 5% C0 2 . 

8. Leave the cultures undisturbed for 7 d, after which time replace the medium with 
an equal volume of fresh medium taking care not to dislodge the explants. 

9. Check for outgrowth of cells at 7-10 d (see Note 8). 

10. Replace the medium at 14 d and twice weekly thereafter until the desired cell 
density has been attained. 

3.2. Passaging Cells and Establishing Secondary Cultures 

1. Remove and discard the spent medium. 

2. Gently wash the cell layers three times with 10 mL of PBS without Ca 2+ and Mg 2+ . 

01/Gall/1-18/F1 8 1 2/26/03, 10:44 AM 

I ■ I I I 


Human Osteoblast Culture 9 

3. To each flask add 5 mL of freshly thawed trypsin-EDTA solution at room tem- 
perature (20°C) and incubate for 5 min at room temperature with gentle rocking 
every 30 sec to ensure that the entire surface area of the flask and explants is 
exposed to the trypsin-EDTA solution. 

4. Remove and discard all but 2 mL of the trypsin-EDTA solution, and then incu- 
bate the cells for an additional 5 min at 37°C. 

5. Remove the flasks from the incubator and examine under the microscope. Look 
for the presence of rounded, highly refractile cell bodies floating in the trypsin- 
EDTA solution. If none, or only a few, are visible tap the base of the flask sharply 
on the bench top in an effort to dislodge the cells. If this is without effect, incu- 
bate the cells for a further 5 min at 37°C. 

6. When most of the cells have become detached from the culture substratum, trans- 
fer to a "universal container" with 5 mL of DMEM with 10% FCS to inhibit 
tryptic activity. 

7. Wash the flask two to three times with 10 mL of SFM and pool the washings with 
the original cell isolate. 

8. Centrifuge at 250g for 5 min to pellet the cells. 

9. Remove and discard the supernatant, invert the tube, and allow the medium to 
drain briefly. 

10. Resuspend the cell in 2 mL of SFM. If the cells are clumping see Note 9. If 
required, the cell suspension can be filtered through a 70-u.m "Cell Strainer" 
(Becton Dickinson) to remove any bone spicules or remaining cell aggregates. 
For convenience and ease of handling the filters have been designed to fit into the 
neck of a 50-mL polypropylene tube. Wash the filter with 2-3 mL of SFM and 
add the filtrate to the cells. 

1 1 . Take 20 uL of the mixed cell suspension and dilute to 80 fxL with SFM. Add5 \jlL 
of trypan blue solution, mix, and leave for 1 min before counting viable (round 
and refractile) and nonviable (blue) cells in a Neubauer Hemocytometer. Using 
this procedure, typically 1-1.5 x 10 6 cells are harvested per 75-cm 2 flask, of 
which a75% are viable. 

12. Plate the harvested cells at a cell density suitable for the intended analysis. We 
routinely subculture at 5 x 10 3 — 10 4 cells/cm 2 and achieve plating efficiencies 
measured after 24 h of a70% {see Note 10). 

3.3. Phenotypic Characterization 

The phenotypic characterization of HBDCs is described in detail in ref. 5. 
The simplest phenotypic marker to investigate is the enzyme alkaline phos- 
phatase, a widely accepted marker of early osteogenic differentiation. Alkaline 
phosphatase can be measured by simple enzyme assay or by histochemical 
staining. Basal activity is initially low, but increases with increasing cell den- 
sity. Treatment with 1 ,25-(OH) 2 D 3 increases alkaline phosphatase activity. The 
most specific phenotypic marker is osteocalcin. This is a protein of Mr 5800 
containing residues of the vitamin K-dependent amino acid y-carboxy glutamic 

01/Gall/1-18/F1 9 I 2/26/03, 10:44 AM 

I ■ I I I 



Table 2 

PCR Primers for a Panel of Osteoblastic Markers 




Primer pairs 

T m (°C) 

Product size (bp) 


5'-ccc tea cac tec teg ccc tat-3' 

5'-tca gec aac teg tea cag tec -3' 



PTH receptor 

5'-agg aac aga tct tec tgc tgc a-3' 

5'-tgc atg tgg atg tag ttg cgc gt-3' 




5'-aag age ttc aaa ccg aga tac aag-3' 


5'-ccg agg ttg gec ccg at-3' 




5'-ccc cac gac aac cgc acc-3' 

5'-cac tec ggc cca caa ate tc-3' 




5'-ggg cgc tac ctt gag ata gag tt-3' 

5'-gag tga cag ttt tgg gaa agt gg-3' 




5'-act att aat gec ace gac atc-3' 

5'-aaa aac tgg ggc tea ate ta-3' 




acid. In humans its synthesis is restricted to mature cells of the osteoblast lin- 
eage. It is an excellent late stage markers for cells of this series despite the fact 
that its precise function in bone has yet to be established. Osteocalcin can be 
measured by one of the many commercially available kits. l,25-(OH) 2 D 3 
increases the production of osteocalcin in cultures of HBDCs, but not fibro- 
blasts obtained from the same donors. More recently, researchers have adopted 
the use of reverse transcription (RT)-PCR to look at the expression of osteo- 
blastic markers in HBDCs. PCR primers for a panel of osteoblastic markers 
including osteocalcin are shown in Table 2. 

3.4. Phenotypic Stability in Culture 

As a matter of routine we perform all of our studies on cells at first passage. 
Other investigators have studied the effects of repeated subculture on the phe- 
notypic stability of HBDCs and found that they lose their osteoblast-like char- 
acteristics. In practical terms this presents real difficulties, as it is often 
desirable to obtain large numbers of HBDCs from a single donor. As an alter- 
native to repeated subculture, trabecular explants can be replated at the end of 
primary culture into a new flask (see Fig. 1). Using this technique, it is possible 
to obtain additional cell populations that continue to express osteoblast-like 
characteristics, including the ability to mineralize their extracellular matrix, 
and maintain their cytokine expression profile (6). Presumably, these cultures 
are seeded by cells that are situated close to the bone surfaces, and that retain 



2/26/03, 10:44 AM 

I ■ I I I 


Human Osteoblast Culture 1 1 

the capacity for extensive proliferation and differentiation. The continued sur- 
vival of these cells may be related to the gradual release over time in culture of 
the cytokines and growth factors that are known to be present in the extracellu- 
lar bone matrix, many of which are known to be produced by mature cells of 
the osteoblast lineage. The addition of 25 \iM L-ascorbic acid (50 fxg/mL) (see 
Note 2) to HBDCs in secondary culture (E1P1) produces a sustained increase 
in the deposition of matrix due to an increase in the synthesis of collagen and 
noncollagenous protein and bone sialoprotein and osteocalcin. 

3.5. Passaging Cells Cultured in the Continuous Presence 
of Ascorbate 

Because of their synthesis and secretion of an extensive collagen-rich extra- 
cellular matrix, HBDCs cultured in the continuous presence of ascorbate can- 
not be subcultured using trypsin-EDTA alone. They can, however, be 
subcultured if first treated with purified collagenase. The basic procedure is as 

1. Rinse the cell layers twice with SFM (10 mL/75-cm 2 flask). 

2. Incubate the cells for 2 h at 37°C in 10 mL of SFM containing 25 U/mL of puri- 
fied collagenase (Sigma type VII) and 2 mM additional calcium (1:500 dilution 
of a filter-sterilized stock solution of 1 M CaCl 2 ). 

3. Gently agitate the flask for 10-15 sec every 30 min. 

4. Terminate the collagenase digestion by discarding the medium (check that there 
is no evidence of cell detachment at this stage). 

5. Gently rinse the cell layer twice with 10 mL of Ca 2+ - and Mg 2+ -free PBS. To 
each flask add 5 mL of freshly thawed trypsin-EDTA solution, pH 7.4, at room 
temperature (20°C). 

6. Typically this procedure yields -3.5-4 x 10 6 cells/75-cm 2 flask after 28 d in pri- 
mary culture. Cell viability is generally >90%. 

3.6. Setting Up Mineralizing HBDC Cultures 

The function of the mature osteoblast is to form bone. Despite the over- 
whelming evidence that cultures of HBDCs contain cells of the osteoblast lin- 
eage, initial attempts to demonstrate the presence of osteogenic (i.e., bone 
forming) cells proved unsuccessful. Subsequently, several authors reported that 
culture of HBDCs in the presence of ascorbate and millimolar concentrations 
of the organic phosphate ester (3-glycerol phosphate ((3-GP) led to the forma- 
tion of mineralized structures resembling the nodules that form in cultures of 
fetal or embryonic animal bone derived cells (reviewed in ref. 22). These have 
been extensively characterized and shown by a variety of morphological, bio- 
chemical, and immunochemical criteria to resemble embryonic/woven bone 
formed in vivo. An alternative to the use of (3-GP is to provide levels of inor- 

01/Gall/1-18/F1 11 I 2/26/03, 10:44 AM 

I ■ I I I 


72 Gallagher 

ganic phosphate sufficient to support the process of cell-mediated mineraliza- 
tion in vitro, and the preferred method when studying HBDCs, is supplementa- 
tion of the culture medium with inorganic phosphate (Pi; see Note 11). The 
protocol for inducing matrix mineralization in cultures of HBDCs is as follows: 

1. Prepare fragments of human trabecular bone as described in Subheading 3.1., 
steps 1-6 (see Note 12). 

2. Culture the washed bone fragments in medium supplemented with 100 \iM 
L-ascorbic acid 2-phosphate and either 200 vM hydrocortisone or 10 vM dexam- 

3. Culture for 4-5 wk until the cells have attained confluence with medium changes 
twice weekly. 

4. When the cells have synthesized a dense extracellular matrix, subculture using 
the sequential collagenase/trypsin-EDTA protocol and plate the cells in 25-cm 2 
flasks at a density of 10 4 viable cells/cm 2 . 

5. After a further 14 d, supplement the medium with 0.01% phosphate solution (see 
Subheading 2.5.4.). 

6. After 48-72 h, wash the cell layers two to three times with 10 mL of SFM. 

7. Fix with 95% ethanol at 4°C (see Note 13). 

3.7. Measuring Alkaline Phosphatase Activity 

1. Add 2.5 mL of staining solution from the Sigma 86-R Staining Kit to each flask 
of cells (or enough to coat the surface). 

2. Place specimens in a humidified chamber and incubate for 1 h at 20°C in the dark. 

3. Wash under running tap water and counterstain the nuclei for 15 sec with hema- 

4. Mineral deposits can be stained using a modification of von Kossa's technique. 

5. Prior to examination, mount sections in DPX and cell layers in flasks covered 
with glycerol. 

3.8. Cryopreservation of HBDC 

If required, HBDCs can be stored frozen for extended periods in liquid 
nitrogen or in ultralow temperature (-135°) cell freezer banks. We use the fol- 
lowing protocol: 

1. Passage the cells using trypsin-ETDA as described in Subheading 3.2., 
steps 1-5. 

2. Pellet the cells by centrifugation at 250g for 5 min and pour off the supernatant. 

3. Resuspend the cell pellet in FCS, adjust to a density of 1-2 x 10 6 cells/mL in a 
volume of 900 |xL and transfer to a cryoampule. 

4. Swirl the ampule in an ice water bath. 

5. Add 100 uT of DMSO gradually while holding the ampule in the iced water. 

6. Close ampules tightly and freeze at 5°C/min to 4°C, followed by l°C/min 
to -80° C in a cell freezing container. 

7. Transfer the cells to liquid nitrogen for long-term storage. 

01/Gall/1-18/F1 12 I 2/26/03, 10:44 AM 

I ■ I I I 

Human Osteoblast Culture 13 

3.9. Thawing Cells 

1. Retrieve the cells from liquid nitrogen and place in a water bath set at 37°C. 

2. Transfer the cells to a universal container to which at least 20 volumes of pre- 
heated culture medium has been added. 

3. Centrifuge at 250g for 2 min to pellet the cells and pour off the supernatant. 

4. Resuspend the cells in approx 10 mL of the medium and place into culture for 24 h. 

5. Replace the medium after 24 h and culture for 2-3 wk. 

4. Notes 

1. Although most investigators have used the original explant method with only 
minor modifications, others have developed alternative techniques for the isola- 
tion and culture of HBDCs. Gehron-Robey and Termine used prior digestion of 
minced bone with clostridial collagenase and subsequent culture of explants in 
medium with reduced calcium concentrations (7). In contrast, Wergedal and 
Baylink have used collagenase digestion to liberate cells directly (24). Marie and 
co-workers have used a method in which explants are first cultured on a nylon 
mesh (20). These alternative methods are described in greater detail in ref. 5. 

2. Beresford and co-workers introduced the more stable analogue L-ascorbic acid 2- 
phsophate (Wako Pure Chemical Industries Ltd.) which does not have to be added 
daily (see ref. 6 for details). 

-{©)- 3. Batches of serum vary in their ability to support the growth of HBDCs. It is 

advisable to screen batches and reserve a large quantity of serum once a suitable 
batch has been identified. HBDCs will grow in autologous and heterologous 
human serum, but as yet no comprehensive studies have been performed to iden- 
tify the effects on growth and differentiation. 

4. The authors have obtained consistent results with plasticware from Sarstedt and Becton 
Dickinson. Smaller flasks or dishes can be used if the amount of bone available is <0.2 g. 

5. Several other assays for osteocalcin are commercially available. 

6. Bone can be stored for periods of up to 24 h at 4°C in PBS or SFM prior to culture 
without any deleterious effect on the ability of the tissue to give rise to popula- 
tions of osteoblastic cells. 

7. Bone cells have also been cultured successfully from many other anatomical sites 
including tibia, femur, rib, vertebra, patella, and digits. 

8. With the exception of small numbers of isolated cells, which probably become 
detached from the bone surface during the dissection, the first evidence of cellu- 
lar proliferation is observed on the surface of the explants, and this normally 
occurs within 5-7 d of plating. After 7-10 d, cells can be observed migrating 
from the explants onto the surface of the culture dish (see Fig. 3). If care is taken 
not to dislodge the explants when feeding, and they are left undisturbed between 
media changes, they rapidly become anchored to the substratum by the cellular 
outgrowths. The typical morphology of the cells is shown in Fig. 4, but cell shape 
varies between donors, from fibroblastic to cobblestone-like. Cultures generally 
attain confluence 4-6 wk post plating, and typically achieve a saturation density 
of 29,000 ± 9000 cells/cm 2 (mean + SD, N = 1 1 donors). 

01/Gall/1-18/F1 13 1 2/26/03, 10:44 AM 

I ■ I I I 



Fig. 3. Migration of cells expressing osteoblastic characteristics (HBDCs) from 
explanted cancellous bone. 


Fig. 4. Typical morphology of cells expressing osteoblastic characteristics 
(HBDCs) from explanted cancellous bone {see Note 14). 

9. If the cells are clumping, resuspend in 2 mL of SFM containing 1 jxg/mL of DNase I 
for each dish or flask treated with trypsin-EDTA, and using a narrow-bore 
2-mL pipet, repeatedly aspirate and expel the medium to generate a cell suspen- 
10. In our experience the minimum plating density for successful subculture is 
3500 cells/cm 2 . Below this the cells exhibit extended doubling times and often 
fail to grow to confluence. Note that cells can be passaged successfully onto a 


2/26/03, 10:44 AM 

I ■ I I I 

Human Osteoblast Culture 



Fig. 5. Human primary derived osteoblasts cultured (7 d) on nanotopography tex- 
tured titanium-coated silicon immunostained for (3-tubulin to visualize the microtubles 
for specific orientations correlating to the stimulus of nanotopography. Image taken 
using a CLSM310 Zeiss confocal microscope. (J. M. Rice, J. A. Hunt and J. A. 
Gallagher, unpublished.) 

range of substrates. Recently HBDC culture has been used to investigate 
biocompatibility and osteogenicity of novel biomaterials. Figure 5 shows HBDCs 
passaged onto nanotopography textured titanium coated silicon and subsequently 
stained with antibodies specific for (3-tubulin. 
1 1 . Mineralizing cultures: HBDCs cultured in the continuous presence of glucocorti- 
coids and the long-acting ascorbate analogue produce a dense extracellular matrix 
that mineralizes extensively following the addition of Pi. This is the case for the 
original cell population (E1P1) and that obtained following replating of the tra- 
becular explants (E2P1), which further attests to the phenotypic stability of the 
cultured cells. Cells cultured in the continuous presence of ascorbate and treated 
with glucocorticoids at first passage show only a localized and patchy pattern of 
mineralization, despite possessing similar amounts of extracellular matrix and 
alkaline phosphatase activity. Cells cultured without ascorbate, irrespective of 
the presence or absence of glucocorticoids, secrete little extracellular matrix, and 
do not mineralize. The ability of the cells to mineralize their extracellular matrix 
is dependent on ascorbate being present continuously in primary culture. The 
addition of ascorbate in secondary culture, even for extended periods, cannot 
compensate for its omission in primary culture. This finding provides further 



2/26/03, 10:44 AM 

I ■ I I I 


16 Gallagher 

evidence to support the hypothesis that maintenance of adequate levels of ascor- 
bate during the early stages of explant culture is of critical importance for the sur- 
vival of cells that retain the ability to proliferate extensively and give rise to 
precursors capable of undergoing osteogenic differentiation. HBDCs cultured con- 
tinuously in the presence of ascorbate and glucocorticoids retain the ability to form 
bone when implanted in vivo within diffusion chambers in athymic mice (23). 

12. For studies of in vitro mineralization, it is preferable to obtain trabecular bone 
from sites containing hematopoietic marrow such as the upper femur or iliac crest. 

13. Fixation of mineralized cultures: This can be done in situ, for viewing en face, or 
if sections are to be cut following detachment of the cell layer from the surface of 
the flask using a cell scraper. Great care is needed if the cell layer is to be har- 
vested intact, particularly when mineralized. 

14. The available evidence indicates that cultures of HBDCs contain cells of the 
osteogenic lineage at all stages of differentiation and maturation. This conclu- 
sion is consistent with the expression of both early (alkaline phosphatase) and 
late (osteocalcin, bone sialoprotein) stage markers of osteoblast differentiation. 
In addition, in ascorbate-treated cultures there is a small subpopulation (s5%) of 
cells that express the epitope recognized by the monoclonal antibody (MAb) 
STRO-1 (25), which is a cell-surface marker for clonogenic, multipotential mar- 
row stromal precursors capable of giving rise to cells of the osteogenic lineage in 
vitro. The presence of other cell types, including endothelial cells and those 
derived from the hematopoietic stem cell, has been investigated using a large 
panel of MAbs and flow cytometry and/or immunocytochemistry. The results of 
these studies reveal that at first passage there are no detectable endothelial, lym- 
phoid, or erythroid cells present. A consistent finding, however, is the presence 
of small numbers of cells (s5%) expressing antigens present on cells of the mono- 
cyte/macrophage series. In the absence of added calcitriol, particularly in SFM or 
FCS that has been depleted of endogenous calcitriol by charcoal treatment, the 
amount of ostecalcin produced by HBDCs is below the limits of detection in 
most assays (3,4). The same applies to the detection of steady-state levels of 
osteocalcin mRNA. An exception to this general rule is when HBDCs are cul- 
tured for extended periods in the presence of L-ascorbate or its stable analog, L- 


I am grateful to Jane Dillon and Paula Finnigan for their assistance in pre- 
paring this chapter and to Dr. John Hunt and Dr. Judith Rice, UKCTE, Clinical 
Engineering, University of Liverpool for Fig. 3. 


1. Bard, D. R., Dickens, M. J., Smith, A. U., and Zarek, J. M. (1972) Isolation of 
living cells from mature mammalian bone. Nature 236, 314-315. 

2. Mills, B. G., Singer, F. R„ Weiner, L. P., and Hoist, P. A. (1979) Long term culture 
of cells from bone affected with Paget's disease. Calcif. Tissue Int. 29, 79-87. 

01/Gall/1-18/F1 16 1 2/26/03, 10:44 AM 

I ■ I I I 


Human Osteoblast Culture 1 7 

3. Gallagher, J. A., Beresford, J. N., McGuire, M. K. B., et al. (1984) Effects of 
glucocorticoids and anabolic steroids on cells derived from human skeletal and 
articular tissues in vitro. Adv. Exp. Med. Biol. 171, 279-292. 

4. Beresford, J. N., Gallagher, J. A., Poser, J. W., and Russell, R. G. G. (1984) Pro- 
duction of osteocalcin by human bone cells in vitro. Effects of l,25(OH)2D3, 
parathyroid hormone and glucocorticoids. Metab. Bone Dis. Rel. Res. 5, 229-234. 

5. Gallagher, J. A., Gundle, R., and Beresford, J. N. (1996) Isolation and culture of 
bone forming cells (osteoblasts) from human bone, in Human Cell Culture Proto- 
cols (Jones, G. E., ed.), Humana Press Totowa, NJ. 

6. Gundle, R., Stewart, K., Screen, J., and Beresford, J. N. (1998) Isolation and cul- 
ture of human bone derived cells, in Marrow Stromal Cell Culture (Beresford, J. 
and Owen, M., eds.), Cambridge University Press, Cambridge, UK. 

7. Gehron Robey, P. and Termine, J. D. (1985) Human bone cells in vitro. Calcif. 
Tissue Int. 37, 453-460. 

8. Auf'mkolk, B., Hauschka, P. V., and Schwartz, R. (1985) Characterisation of 
human bone cells in culture. Calcif. Tissue Int. 37, 228-235. 

9. MacDonald, B. R., Gallagher, J. A., Ahnfelt-Ronne, I., Beresford, J. N., Gowen, 
M., and Russell, R. G. G. (1984) Effects of bovine parathyroid hormone and 
l,25(OH)2D3 on the production of prostaglandins by cells derived from human 
bone. FEBS Lett. 169, 49-52. 

10. MacDonald, B. R., Gallagher, J. A., and Russell, R. G. G. (1986) Parathyroid 
hormone stimulates the proliferation of cells derived from human bone. Endocri- 
nology 118, 2245-2449. 

11. Beresford, J. N., Gallagher, J. A., and Russell, R. G. G. (1986) 1,25-Dihydroxy- 
vitamin D 3 and human bone derived cells in vitro: effects on alkaline phosphatase, 
type I collagen and proliferation. Endocrinology 119, 1776-1785. 

12. Fedarko, N. S., Vetter, U., Weinstein, S., and Robey, P. G. (1992) Age-related 
changes in hyaluronan, proteoglycan, collagen and osteonectin synthesis by 
human bone cells. /. Cell Physiol. 151, 215-227. 

13. Beresford, J. N., Gallagher, J.A., Gowen, M., et al. (1984) The effects of mono- 
cyte-conditioned medium and interleukin 1 on the synthesis of collagenous and 
non-collagenous proteins by mouse bone and human bone cells in vitro. Biochim. 
Biophysica. Acta Gen. Subj. 801, 58-65. 

14. Gowen, M., Wood, D. D., and Russell, R. G. (1985) Stimulation of the prolifera- 
tion of human bone cells in vitro by human monocyte products with interleukin- 1 
activity in/. Clin. Invest. 4, 1223-1229. 

15. Vaishnav, R., Gallagher, J. A., Beresford, J. N., Poser, J. W., and Russell, R. G. 
G. (1984) Direct effects of stanozolol and oestrogen on human bone cells in cul- 
ture, in Osteoporosis. Proceedings of Copenhagen International Symposium, pp. 

16. Eriksen, E. F., Colvard, D. S., Berg, N. J., et al. (1988) Evidence of estrogen 
receptors in normal human osteoblast-like cells. Science 241, 84-86. 

17. Schoefl, C, Cuthbertson, K. S. R., Walsh, C. A., et al. (1992) Evidence for P2- 
purinoceptors on osteoblast-like cells. /. Bone Min. Res. 7, 485-591. 

01/Gall/1-18/F1 17 I 2/26/03, 10:44 AM 

I ■ I I I 


18 Gallagher 

18. Bowler, W. B., Gallagher, J. A., and Bilbe, G. (1995) Identification and cloning 
of human P2U purinceptor present in osteoclastoma, bone and osteoblasts. /. Bone 
Miner. Res. 10, 1137-1145. 

19. Ralston, S. H., Todd, D., Helfrich, M., Benjamin, N., and Grabowski, P.S. (1994) 
Human osteoblast-like cells produce nitric oxide and express inducible nitric oxide 
synthase. Endocrinology 135, 330-336. 

20. Marie, P. J., Sabbagh, A., De Vernejoul, M. C, and Lomri, A. (1988) Osteocalcin 
and deoxyribonucleic acid synthesis in vitro and histomorphometric indices of 
bone formation in postmenopausal osteoporosis. /. Clin. Invest. 69, 272-279. 

21. Walsh, C. A., Birch, M. A., Fraser, W. D., et al. (1995) Expression and secretion 
of parathyroid hormone-related protein by human osteblasts in vitro: effects of 
glucocorticoids. /. Bone Miner. Res. 10, 17-25. 

22. Beresford, J. N., Graves, S. E., and Smoothy, C. A. (1993) Formation of 
mineralised nodules by bone derived cells in vitro: a model of bone formation? 
Am. J. Med. Genet. 45, 163-178. 

23. Gundle, R. G., Joyner, C. J., and Triffitt, J. T. (1995) Human bone tissue forma- 
tion in diffusion chamber culture in vivo by bone derived cells and marrow stro- 
mal cells. Bone 16, 597. 

24. Wergedal, J. E. and Baylink, D. J. (1984) Characterisation of cells isolated and 
cultured from human trabecular bone. Proc. Soc. Exp. Biol. Med. 176, 60-69. 

25. Walsh, S., Jefferiss, C, Stewart, K., Jordan, G. R., Screen, J., and Beresford, J. N. 
(2000) Expression of the developmental markers STRO-1 and alkaline phos- 
phatase in cultures of human marrow stromal cells: regulation by fibroblast growth 
factor (FGF)-2 and relationship to the expression of FGF receptors 1-4. Bone 27, 

01/Gall/1-18/F1 18 1 2/26/03, 10:44 AM 

I ■ I I I 


Osteoblast Isolation from Murine Calvariae 
and Long Bones 

Astrid Bakker and Jenneke Klein-Nulend 

1. Introduction 

When conducting in vitro research on bone, a choice has to be made between 
using bone organ or bone cell cultures. When one decides to use the latter, the 
question is whether to use primary cells or cell lines. The advantage of using 
cell lines over freshly isolated cells lies in the ready availability of large num- 
bers of cells, the homogeneity of the cell cultures, and the expected invariabil- 
ity of the phenotype. In the long run, however, cell lines appear unstable to 
some extent. In addition, their clonal selection has favored rapidly growing 
cells, but has not necessarily selected for the whole range of bone-specific gene 
expression characteristic of primary bone cells. This means that in certain 
experiments, the use of primary bone cells is preferred to the use of cell lines. 

Peck and co-workers initiated the use of primary bone cell cultures in 1964 
(1). They isolated cells from frontal and parietal bones of fetal and neonatal rat 
calvariae by collagenase digestion of the uncalcified bone matrix. The isolated 
cells were viable, proliferated during culture, and exhibited high activity of the 
osteoblast marker alkaline phosphatase (ALP). The real nature of the cells, 
however, especially the amount of contamination with connective tissue fibro- 
blasts, could not be defined unambiguously (1). Wong and Cohn (1974) tried 
to isolate a better defined and more homogeneous cell population by removing 
the outer layers of the periosteum with successive collagenase treatments (2). 
Although this method led to cell cultures that were more osteoblastic in nature, 
these were not free from other cell types, such as osteoclast precursors, either 
(3). Other investigators have tried to improve the osteoblastic character of the 
isolated bone cell populations by removing the fibroblastic outer periosteum 
before using enzymatic digestion to isolate the cells from the calvarium (4,5). 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


02/Nulend/19-28/F1 19 1 2/26/03, 10:44 AM 

I ■ I I I 


20 Bakker and Klein-Nulend 

This method resulted in two cell populations, one of which was still osteogenic 
after prolonged culture time (osteoblastic cells), and another one that was not 
(periosteal fibroblasts) (6). These studies led to a broad range of methods for 
obtaining well-defined osteoblast-like cells in vitro, which are at present widely 
used as a tool to improve our knowledge of bone biology (7-9). 

This chapter describes the isolation of primary mouse bone cells from adult 
mouse calvariae and long bones, as well as the process of isolation of bone 
cells from neonatal mouse calvariae. Owing to their difference in origin and 
method of isolation, it is to be expected that each of the primary bone cell 
cultures described will have its own characteristics. For example, it has been 
shown that neonatal cells show a higher basal release of nitric oxide and a 
higher response to 1,25-dihydrogxyvitamin [D 3 l,25-(OH) 2 D 3 ] treatment than 
bone cells obtained from adult bone (10). Because vitamin D3 stimulates 
immature bone cell differentiation, this supports the notion that neonatal cell 
cultures contain more immature, rapidly growing cells than cultures from adult 
bone. Thus, for in vitro studies investigating the cellular behavior of adult bone, 
it seems advisable to use cells from adult bone fragments to reproduce best the 
inherent cellular properties of the adult tissue. 

Another example relates to mechanosensitivity of bone cells. Because the 
prevailing mechanical strains in the skull are much lower than those in the 
axial and appendicular skeleton, the question has arisen whether cells from 
calvariae or long bones should be used in such studies. We addressed this 
issue by studying the nitric oxide production of the bone cells in response to 
mechanical stimulation in the form of fluid flow. We found no difference in the 
responsiveness of osteoblasts from adult mouse calvariae or adult mouse long 
bones (12). These results suggest that the cellular mechanosensitivity of 
calvariae and long bone cells is not intrinsically different and that either cell 
culture can be used for these sorts of experiments. 

2. Materials 

2.1. Tissues 

1. Cells are obtained from the long bones and the calvariae of adult (age 9 wk or 
older) mice, or the calvariae from neonatal mice pups (age 3-4 d). 

2.2. Instruments 

All of the following materials have to be sterile. 

1. Polystyrene plate and needles for fixing the mice. 

2. Scalpels (no. 10 and 1 1), scissors, tweezers, and curved forceps. 

3. 5-mL and 10-mL syringes, 27G1/2 needles, disposable cell scrapers, and 0.2-jxm 
disposable filter units. 

02/Nulend/19-28/F1 20 1 2/26/03, 10:44 AM 

I ■ I I I 


Bone Cells from Calvariae and Long Bones 21 

4. 25-cm 2 tissue culture flasks (Nunc), six-well tissue plates (Costar), 94/16-mm 
cellstar Petri dishes (Greiner), and 145/20-mm cellstar (large) Petri dishes 

5. 100 x 16 mm (10 mL) conical base test tubes with screw cap (Bibby SterilinLtd., 
Staffordshire, UK). 

2.3. Media and Solutions 

1. Phosphate-buffered saline (PBS): 137 mMNaCl, 1.5 mMKH 2 P0 4 , 2.7 mMKCl, 
and 8.1 mM Na 2 HP0 4 . Adjust the pH to 7.4. 

2. Dulbecco's modified Eagle's medium (DMEM; Gibco, Paisley, UK): Add 2.2 g 
NaHC0 3 /L; adjust the pH to 7.4. 

3. Complete culture medium (cCM): DMEM, supplemented with 100 U/mL of peni- 
cillin (Sigma), 50 fxg/mL of streptomycin sulfate (Gibco), 50 u,g/mL of 
gentamycin (Gibco), 1.25 |xg/mL of fungizone (Gibco), 100 ng/mL of ascorbate, 
and 10% fetal bovine serum (FBS) (Hyclone, Logan, UT, USA; see Note 1). 
Make fresh and filter sterilize. 

4. Collagenase solution: 2 mg of collagenase II (Sigma) per milliliter of DMEM. 
Make fresh and filter sterilize. 

5. Trypsin solution: 0.25% trypsin 1:250 (Difco, Detroit, MI, USA) and 0.10% 
EDTA in PBS; filter sterilize. 

6. Digestion solution: Add 1 mL of trypsin solution and 3.2 mg of collagenase II to 
4 mL of PBS. Make fresh. 

7. Vitamin D medium (VDM): DMEM, supplemented with 100 U/mL of penicillin, 
50 |xg/mL of streptomycin sulfate, 50 u.g/mL of gentamicin, 1.25 \xg/mL of 
fungizone, 100 ng/mL of ascorbate, 0.2% bovine serum albumin (Sigma) and 10~ 8 M 
l,25-(OH) 2 D 3 . Make fresh and shield away from direct light. 

8. Vitamin D control medium (VDCM): Composition is the same as vitamin D 
medium, except that the 1 ,25-(OH) 2 D 3 is replaced by an equal amount of vehicle. 

9. BCA Protein Assay Reagent Kit (Pierce, Rockford, IL, USA). 

3. Methods 

Normal techniques for working under sterile conditions (use of sterile media 
and instruments and working in a flow cabinet) should be used to keep the cell 
cultures sterile. 

3. 1. Isolation and Culture of Primary Bone Cells from Adult 
Mouse Long Bones 

1. Euthanize one or two adult mice by means of cervical dislocation. 

2. Fix the mouse in a supine position on a polystyrene plate or in a large Petri dish, 
and clean the abdomen and extremities using 70% ethanol. 

3. Make a single incision through the skin, starting at the top of the sternum and 
ending a few millimeters above the genitals, using a no. 10 scalpel. Make a sec- 
ond incision starting from the top of the first incision and ending at the wrist of 

02/Nulend/19-28/F1 21 I 2/26/03, 10:44 AM 

I ■ I I I 


22 Bakker and Klein-Nulend 

the upper left extremity. Repeat this procedure with the other paws. Carefully 
remove the skin from the abdomen with a blade. 

4. Change your blade for a sterile no. 1 1 scalpel. Remove the muscles from the long 
bones in the limb (femur; tibia and fibula; or humerus, radius, and ulna), and 
scrape the bone with a scalpel until it is clean (see Note 2). Excise the long bone 
and place it in a Petri dish with PBS. 

5. When all the long bones have been removed, cut off the epiphyses. 

6. Flush out the bone marrow with PBS, using a 5-mL syringe and a 27-gauge needle. 

7. Cut the clean diaphyses into small pieces of approx 1-2 mm 2 using scissors. 

8. Wash the bone pieces with PBS, and incubate in 4 mL of collagenase solution at 
37°C in a shaking water bath to remove all remaining soft tissue and adhering 

9. After approx 1 h, vigorously shake the solution by hand. 

10. After 2 h, add 4 mL of cCM containing 10% FBS to inhibit further collagenase 
activity, and rinse the bone pieces three times with cCM. 

1 1 . Transfer the bone pieces to 25-cm 2 flasks, containing 5 mL of cCM, at a density of 
about 20-30 fragments per flask. Replace culture medium three times per week. 

12. Adult mouse bone cells will start to migrate from the bone chips after 3-5 d. On 
average the cell monolayer growing from the bone fragments will reach 
confluency after 11-15 d. 

13. To obtain more cells, trypsinize the monolayer by incubating the cells with 1 mL of 
trypsin solution at 37°C for 10 min. 

14. Plate the cells at 25 x 10 3 cells per well in six-well culture dishes containing 
3 mL of cCM per well. 

15. Change medium three times per week, and after approx 7-10 d cells will reach 
subconfluency, upon which they can be used for experiments (see Note 3 and 
Subheading 3.4.). The average number of cells thus obtained lies between 
4x 10 6 and6x 10 6 cells. 

3.2. Isolation and Culture of Primary Bone Cells from Adult Mouse 

1 . Euthanize two adult mice and fix them on a polystyrene plate, or in a large Petri dish. 

2. Clean the head using 70% ethanol, and make a cut through the skin at the base of 
the skull, using scissors. 

3. Make an incision starting at the nose bridge, and ending at the base of the skull. 
Remove the skin from the top of the head (see Fig. 1). 

4. Use scissors to cut through the bone at the base of the neck. Cut the calvariae 
loose, while holding the head with curved forceps placed in the orbita. 

5. Transfer the calvariae to a Petri dish with PBS and remove the soft tissues using 
tweezers or by scraping with a knife (see Note 2). 

6. Remove the sutures, using scissors, and chop the remaining bone into small frag- 
ments of approx 1-2 mm 2 . 

7. Incubate the fragments for 30 min in 4 mL of collagenase solution at 37° in a 
shaking water bath. 

02/Nulend/19-28/F1 22 I 2/26/03, 10:44 AM 

I ■ I I I 

Bone Cells from Calvariae and Long Bones 


Fig. 1. Schematic presentation of mouse calvariae. (A) Make a cut through the skin 
at the base of the skull, using scissors (short dashed line). For adult tissue, make an 
incision starting at the nose bridge, ending at the base of the skull (long dashed line), 
and remove the skin from the top of the head (arrows). Use scissors to cut away the 
skin from the top of the neonatal mouse heads. (B) Dissect calvariae as indicated by 
area C and remove as much soft tissue as possible. Do not include the shaded area near 
the neck, as this will result in heavy fibroblast contamination of the cultures. Sf, Skin 
flap; C, Calvaria. 





Remove the collagenase solution and replace with fresh collagenase solution. Incu- 
bate another 30 min, then replace the collagenase solution for trypsin solution. 
Incubate in trypsin for 30 min. Replace by 4 mL of collagenase solution for the 
fourth and final incubation step of 30 min. 

Add 4 mL of cCM to the collagenase to inhibit collagenase activity. Rinse the 
bone pieces three times with cCM. 

11. Transfer the bone pieces to 25 cm 2 flasks, containing 5 mL of cCM, at a density 
of approx 20-30 fragments per flask. 

Change the medium three times per week. Adult mouse bone cells will start to 
migrate from the bone chips after 3-5 d. On average the cell monolayer growing 
from the bone fragments will reach confluency after 11-15 d, upon which the 
monolayer is trypsinized by incubating the cells with 1 mL of trypsin solution at 
37°C for 10 min. 

Plate the cells at 25 x 10 3 cells per well in six-well culture dishes containing 
3 mL of cCM per well. 

14. After approx 7-10 d cells will reach subconfluency, upon which they can be used 
for experiments (see Note 3 and Subheading 3.4.)- The average number of cells 
thus obtained lies between 4 x 10 6 and 6 x 10 6 cells. 

3.3. Isolation and Culture of Bone Cells from Neonatal Mouse 

1. Euthanize 20-30 neonatal mice pups (2-3 liters) by decapitation or halothane 
inhalation, and place the heads in a Petri dish with PBS (see Note 4). 

2. Grasp the head by the nape of the neck, and cut the skin away using scissors. 




2/26/03, 10:44 AM 

I ■ I I I 


24 Bakker and Klein-Nulend 

3. Hold the head with curved forceps placed through the orbita, cut the calvariae 
loose along the edge, and place it in a Petri dish with PBS (see Fig. 1 for a dia- 
gram illustrating how to dissect the calvariae). 

4. Pin the calvariae down with tweezers and cut away the edges and sutures with a 
small scalpel. Transfer the calvariae halves to a 25-mL tube with PBS and wash 
twice with PBS. 

5. Incubate the calvariae in 4 mL of digestion solution at 37°C in a shaking water 
bath. After 10 min shake the calvariae by hand for a few seconds. 

6. Incubate for a total of 20 min, and then transfer the supernatant, containing cells, 
to a 10-mL tube. Add 700 \xL of fetal calf serum (FCS) to the cell suspension to 
inhibit collagenase and trypsin activity. 

7. Wash the calvariae with 3 mL of DMEM (without FBS!), shake well, and add the 
supernatant to the tube containing the cell suspension. This is population no. 1. 

8. Add new digestion solution to the calvariae, and repeat the previous three steps 
to obtain population no. 2. During the 20 min that the calvariae have to incubate 
in the water bath, centrifuge cell population no. 1 for 5 min at 300g Discard the 
supernatant, resuspend the cell pellet in 1 mL of cCM, and add to 17 mL of cCM. 
Pipet in a six-well plate at 3 mL of cell suspension per well. 

9. Repeat the entire procedure for a total of four times, to obtain population 
nos. 1-4. 

10. The culture medium is changed 1 d after isolation of the bone cells. 

1 1 . Within approx 5 d cells will reach subconfluency , upon which they are trypsinized 
by incubation with 200 \xL of trypsin solution per well, at 37°C for 10 min. 

12. Population nos. 1 and 2, resembling osteoblast progenitor cells, are pooled, as 
well as population nos. 3 and 4. This latter pooled cell population is enriched 
with cells exhibiting biochemical characteristics associated with differentiated 
osteoblasts, such as high ALP activity and osteopontin expression. Both pooled 
populations can be used directly for experiments (see Note 5). 

13. The number of cells obtained using this method varies between 6 x 10 6 and 
10 x 10 6 cells- see Note 5 for an alternative method for osteoblast isolation from 
neonatal calvariae. 

3.4. Characterization of the Osteoblast Phenotype 
by Determination of ALP Activity 

Primary bone cell cultures are not 100% pure and may contain some fibro- 
blasts and other nonosteoblastic cell types. It is therefore advisable to check 
routinely the osteoblastic phenotype of the cultures. Because vitamin D 3 stimu- 
lates the differentiation of immature bone cells, leading to enhanced ALP 
activity (11), the osteoblastic phenotype of the primary mouse bone cell cul- 
tures can be determined as follows: 

1. Take a subconfluent cell culture in a six-well plate, wash the cells once with 
PBS, and replace the medium by 3 mL of either VDM or VDCM per well. After 
3 d of incubation, remove the medium and wash the cell layer with PBS. 

02/Nulend/19-28/F1 24 I 2/26/03, 10:44 AM 

I ■ I I I 


Bone Cells from Calvariae and Long Bones 25 

2. Put the cells on ice and add 1 mL of cold milliQ water to the cells. Harvest the 
cells with a cell scraper, and transfer the cell suspension to a 10-mL tube. Soni- 
cate on ice for 10 min, and then centrifuge for 10 min at 500 g. Transfer the 
supernatants for determination of ALP activity and total protein content. 

3. Determine ALP activity by using p-nitrophenyl phosphate as a substrate at pH 10.3, 
according to the method described by Lowry (12). Read the absorbance at 405 nm. 

4. Measure the protein content of the homogenate using a BCA Protein Assay 
Reagent Kit according to the manufacturer's protocol. Read the absorbance at 
570 nm. On average the incubation of adult mouse bone cell cultures with 1,25- 
(OH) 2 D 3 will result in a twofold induction of ALP production. In neonatal mouse 
calvarial cultures the induction of ALP production by l,25-(OH) 2 D 3 is on aver- 
age sixfold. 

4. Notes 

1. Addition of serum to the medium is necessary for the survival and stimulation of 
proliferation of the primary mouse bone cells. "Serum" is not a constant and 
homogeneous product, however, and the growth rate of primary bone cells can 
vary considerably between several batches of serum. It is therefore recommended 
to test several batches of serum on their cell proliferative ability and continue to 
use the one that produces the best results. 

2. Sometimes the primary bone cultures can contain fibroblasts, which grow faster 
than the bone cells and can quickly overgrow the primary bone cell cultures. If 
this problem occurs care should be taken to remove all soft tissues better by scrap- 
ing the bones with a knife before starting the collagenase treatment. Also make 
sure the collagenase is not expired, and that the collagenase solution is made 
fresh every time. 

3. The exact nature of the bone cells that are isolated from adult long bones and 
adult calvariae has not yet been determined. Because the cell isolation protocols 
involve removing the soft tissues and all adhering cells by means of incubation 
with collagenase, the cells that are isolated from the bones might represent osteo- 
cytes that reverted to proliferation after several days of exposure to fetal calf 
serum. The appearance of the isolated bone cells, however, is mostly osteoblastic 
(Fig. 2), and several osteoblast specific markers are expressed by these cells. 
Absence of staining for von Willebrand factor (factor VIII) shows that the bone 
cell cultures do not contain endothelial cells. 

4. Smaller numbers of calvariae can be used successfully, leading to a proportion- 
ately lower yield of osteoblasts. 

5. We prefer the use of collagenase type II; however, other groups have 
reported use of crude collagenase type IA (Sigma C9891), which is cheaper 
and intrinsically contains trypsin as a contaminant. An alternative protocol 
for isolation of osteoblasts from neonatal mouse calvariae, which uses 
alternate collagenase and EDTA incubations to remove as much mineral- 
ized matrix as possible and increase cellular yield, is given below in the 
following subheadings. 

02/Nulend/19-28/F1 25 1 2/26/03, 10:44 AM 

I ■ I I I 


Bakker and Klein-Nulend 


Fig. 2. Phase contrast microscopy of primary mouse bone cell cultures. (A) Adult 
mouse bone cells growing out of the bone chips, d 6 of culture. (B) Subconfluent layer 
of adult mouse bone cells, first passage. (C) Neonatal mouse calvarial cells, popula- 
tion nos. 1 and 2, d 3 of culture. (D) Neonatal mouse calvarial cells, population nos. 3 
and 4, d 3 of culture. Note the cuboidal morphology of the osteoblasts. 


1. Stock solution of collagenase type I at 10 mg/mL in Hanks' balanced salt solu- 
tion (HBSS, Gibco). Filter sterilize and freeze aliquots for single use at -20°C. 
Dilute in HBSS to 1 mg/mL just before use. HBSS contains calcium, which 
increases the activity of the collagenase. 

2. 4 mM EDTA in PBS without calcium and magnesium. Filter sterilize and store at 4 C C. 

3. Conical 15-mL centrifuge tubes (polypropylene, to reduce cell loss by minimiz- 
ing adhesion to tube). Conical 25-mL centrifuge tubes for incubations. Sterile 
Petri dishes for dissection (ideally glass). 

4. Sterile small curved forceps and spring bow scissors. 


1 . Dissect calvariae as described in Subheading 3.3., steps 1-3 and collect them in HBSS. 

2. All incubations are carried out in 3 mL of solution (making sure calvaria are 
completely covered) in a 25-mL centrifuge tube in a shaking water bath at 37°C. 



2/26/03, 10:44 AM 

I ■ I I I 


Bone Cells from Calvariae and Long Bones 27 

3. Incubate calvariae in 1 mg/mL of collagenase for 10 min (fraction 1). 

4. Replace the collagenase solution with fresh solution and discard fraction 1. 

5. Incubate for 30 min in collagenase. Collect the cell suspension (fraction 2) and 
place in a conical centrifuge tube. Wash calvariae in 7 mL of PBS and add wash 
to fraction 2. 

6. Add EDTA solution and incubate for 10 min. Collect the resulting cell suspen- 
sion (fraction 3). Wash calvariae in 7 mL of HBSS and add wash to fraction 3. 

7. Add collagenase and incubate for 30 min (fraction 4). Wash in HBSS and add 
wash to fraction 4. 

8. Further fractions can be collected by repeating steps 5 and 6, but cell yields will 
be increasingly lower. 

9. Centrifuge all fractions immediately after collection (250g for 5 min) and resus- 
pend pellets in cCM (see Subheading 2.3.3.). 

10. Plate out pooled or single fractions in 75-cm 2 culture flasks using cells derived 
from two or three animals per flask. Later fractions will contain the most differ- 
entiated cells. Cells isolated in this way (pooled fractions 2-4) have been used 
successfully in cocultures to generate osteoclasts (see Chapter 1 1 by van 't Hof, 
this volume). 

1 1 . Cultures will be confluent in 3^1 d. To minimize contamination by other adher- 
ent cell types, replace medium once cells have adhered (2-3 h after plating). 


1. Peck, W. A., Birge, S.J., and Fedak, S. A (1964) Bone cells: biochemical and 
biological studies after enzymatic isolation. Science 146, 1476-1477. 

2. Wong G. L. and Cohn, D. V. (1974) Separation of parathyroid hormone and calci- 
tonin-sensitive cells from non-responsive cells. Nature 252, 713-715. 

3. Burger, E. H., Boonekamp, P. M., and Nijweide, P. J. (1986) Osteoblast and 
osteoclast precursors in primary cultures of calvarial bone cells. Anat. Rec. 214, 

4. Yagiela, J. A. and Woodbury, D. M. (1977) Enzymatic isolation of osteoblasts 
from fetal rat calvaria. Anat. Rec. 188, 287-306. 

5. Nijweide, P. J., van der Plas, A., and Scherft, P.J. (1981) Biochemical and histo- 
logical studies on various bone cell preparations. Calcif. Tissue Int. 33, 529-540. 

6. Nijweide, P.J., van Iperen-van Gent, A.S., Kawilarang-de Haas, E. W. M., van der 
Plas, A., and Wassenaar, A.M. (1982) Bone formation and calcification by iso- 
lated osteoblast-like cells. /. Cell Biol. 93, 318-323. 

7. Klein-Nulend, J., Burger, E. H., Semeins, C. M., Raisz, L. G., and Pilbeam C. C. 
(1997) Pulsating fluid flow stimulates prostaglandin release and inducible pros- 
taglandin G/H synthase mRNA expression in primary mouse bone cells. /. Bone 
Miner. Res. 12,45-51. 

8. Klein-Nulend, J., Semeins, C. M., Ajubi, N. E., Nijweide, P. J., and Burger, E. H. 
(1995) Pulsating fluid flow increases nitric oxide (NO) synthesis by osteocytes 
but not periosteal fibroblasts: correlation with prostaglandin upregulation. 
Biochem. Biophys. Res. Commun. 217, 640—648. 

02/Nulend/19-28/F1 27 I 2/26/03, 10:44 AM 

I ■ I I I 


Bakker and Klein-Nulend 

9. Bakker, A. D., Soejima, K., Klein-Nulend, J., and Burger, E. H. (2001) The pro- 
duction of nitric oxide and prostaglandin E2 by primary bone cells is shear stress 
dependent. /. Biomech. 34, 671-677 '. 
10. Soejima, K., Klein-Nulend, J., Semeins, C. M., and Burger, E. H. (2001) Differ- 
ent responsiveness of cells from adult and neonatal mouse bone to mechanical 
and biochemical challenge. /. Cell Phys. 186, 366-370. 

Auf 'mkolk, B., Hauschka, P. V., and Schwartz, E.R. (1985) Characterization of 
human bone cells in culture. Calcif. Tissue Int. 37, 228-235. 
Lowry, O. H. (1955) Micromethods for the assay of enzyme. II Specific pocedure. 
Alkaline phosphatase. Methods Enzymol. 4, 371. 






2/26/03, 10:44 AM 

I ■ I I I 


Mineralizing Fibroblast-Colony-Forming Assays 

Andrew Scutt, Lynsey Reading, Nanette Scutt, and Karen Still 

1. Introduction 

Bone formation does not lend itself easily to investigation because bone 
tissue consists of various cell types embedded in a complex extracellular 
matrix. These cells interact with each other and with the extracellular matrix, 
and when cell populations are removed from the network they cease to func- 
tion normally. In the past, bone cell differentiation was studied using histologi- 
cal methods in either whole embryos or organ cultures. Although this has 
provided much information regarding the temporal and spatial relationships of 
the various cells, the complexity of organ culture systems does not easily allow 
one to investigate the molecular mechanisms involved in bone development 
and mineralization. Cell culture techniques have given us much information 
regarding the mechanistic aspects of gene regulation and cell signaling in 
osteoblastic cells, but isolated osteoblasts do not respond to exogenous agents 
in a similar manner to that observed in vivo (1). Recently a number of in vitro 
models have been established that re-create discrete elements of the cellular 
network present in the bone micro-environment. The advantage of these mod- 
els is that they have reduced complexity compared with organ cultures yet 
retain osteoblasts and their progenitors at various stages of differentiation, 
allowing defined aspects of bone formation to be investigated at the cellular 
and molecular levels. These models are modifications of nodule cultures or 
fibroblast-colony-forming unit (CFU-f) cultures. In CFU-f cultures, bone mar- 
row cells are cultured at relatively low densities under conditions that allow 
the individual CFU-f to adhere and proliferate to form colonies. Because of the 
low plating density, the colonies grow essentially in isolation. Each colony 
therefore represents the clonal expansion of one CFU-f and the assay is consid- 
ered to be a measure of the number of CFU-f present in the original bone 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


03/Scutt/29-40/F1 29 I 2/26/03, 10:44 AM 



Scutt et al. 








Fig. 1. Effect of bFGF on fibroblastic colony formation. By sequentially staining 
the cultures for alkaline phosphatase, calcium, collagen, and total colonies, it can be 
seen that bFGF stimulates not only colony formation but also the differentiation pro- 
cess. Analysis shows that the percentage of colonies positive for collagen and calcium 
increases from about 10% to 65% after treatment with 10 ng/mL of bFGF whereas the 
percentage ALP-positive colonies remains constant at about 80%. 

marrow cell suspension (for review see ref. 2). Given the appropriate cul- 
ture conditions, a proportion of these colonies will differentiate and de- 
velop osteoblastic characteristics such as the expression of alkaline 
phosphatase, collagen accumulation, and calcification. The colonies that 
have all three osteoblastic characteristics are considered to be derived from 
osteoprogenitor cells (3). CFU-f cultures also respond to many agents that 
stimulate bone formation in vivo. For example, in Fig. 1 it can be seen that 
10 ng/mL of basic fibroblast growth factor (bFGF) stimulates not only total 
colony number but also the number of differentiated colonies. In nodule 
cultures the cells are plated at relatively high density, reach confluence, 
and subsequently form three-dimensional structures (nodules) that possess 
many bone-like characteristics (4-6). Although the number of nodules has 
been shown to be dependent on the initial number of osteoblast progenitors 
(7), there exists the possibility of considerable interaction between the vari- 
ous cell types in the cultures owing to the high cell density. This chapter 
deals exclusively with CFU-f cultures as these have been used with some 
success to elucidate the role of bone marrow osteoprogenitor cells in bone 
formation in vitro (8-12) and ex vivo (13-20). 



2/26/03, 10:44 AM 

I ■ I I I 


Mineralizing Fibroblast-Colony-Forming Assays 31 

2. Materials 

2. 1. Isolation of Bone Marrow Cells 

1. Assorted sterile dissection instruments including bone cutters, scissors, scalpels, 
and forceps. 

2. Rotary saw or similar. In our laboratory we use a "Hobby drill" with a cutting 
wheel attachment. 

3. Sterile 1.5-mL microfuge tubes. 

4. Plastic supports cut from either pipet tips, 0.5-mL microfuge tubes, or hypoder- 
mic needle sleeves. These should be cut to size so that they fit inside the 
Eppendorf tubes and raise the bones 3-4 mm from the base of the tubes. 

2.2. Tissue Culture Medium 

Dulbecco's modified Eagle's medium (DMEM) containing 4500 mg/L of 
glucose, 1 mM of pyruvate, 2 mM Glutamax, 50 U/L of penicillin, 50 [xg/L of 
streptomycin, 50 ^ig/mL of ascorbic acid, 10" 8 M dexamethasone and 10% fe- 
tal calf serum (FCS) (see Note 1). 

2.3. Staining and Destaining 

1. Fixative: 70% ethanol (see Note 2). 

2. Alkaline phosphatase (ALP) staining solution: 20 mM Tris, pH 7.5, containing 
0.5 mg/mL of naphthol phosphate AS-BI and 1 mg/mL of fast red B. Prepare 
fresh before use (see Note 3). 

3. ALP destaining solution: 100% Ethanol. 

4. Alizarin red solution: To assess mineralization. 1 mg/mL of alizarin red in dis- 
tilled water adjusted to pH 5.5 with ammonium hydroxide. 

5. Alizarin red destaining solution: 5% perchloric acid in distilled water. 

6. Collagen staining solution: To assess collagen production. Add 50 mg of sirius 
red to 50 mL of saturated picric acid. 

7. Collagen destaining solution: 0. 1 N NaOH mixed with methanol (50:50). Prepare 
fresh just before use. 

8. Borate buffer. 10 mM boric acid; adjust pH to 8.8 with NaOH. 

9. Total-colony staining solution: 1 mg/mL of methylene blue in 10 mM borate 
buffer, pH 8.8. 

2.4. Colony Quantitation 

1. Kodak DC50 digital camera or similar. 

2. Good quality white light transilluminator. 

3. Adobe Photoshop or similar image editing software. 

4. Intelligent quantifier or similar colony counting software. 

3. Methods 

3. 1. Isolation of Whole Bone Marrow Cells 

1 . Use four 125-200-g male Wistar rats and euthanize these by cervical dislocation (see 
Note 4). 

03/Scutt/29-40/F1 31 1 2/26/03, 10:44 AM 

I ■ I I I 


Scutt et al. 


Fig. 2. Recovery of tibial and femoral bone marrow cells using a centrifugation 
method. Bone marrow cells are isolated using a brief centrifugation step. Excised and 
prepared bones are placed in 1.5-mL microfuge tubes supported by plastic inserts. The 
bone marrow cells are expelled by centrifuging the samples at 2000 rpm in a microfuge 
for 5 sec. The bone marrow pellet is then resuspended in 10 mL of culture medium as 
described in Subheading 3. 

2. Remove the tibias and femurs from each animal under aseptic conditions and 
dissect away the soft tissues. For the tibia this can be best achieved by grasping 
the growth plate with bone cutters and tearing the growth plate off together with 
any attached musculature. The muscle can then be removed cleanly and the bone 
cut off at the junction of the tibia and fibula using a rotary saw. The condyle of 
the femur can be grasped and torn off in a similar manner. However, not all of the 
muscle will come off cleanly and will have to be removed carefully using either 
scissors or a scalpel. The bone can then be cut as close to the femoral neck as 
possible using a rotary saw. 

3. Recover the bone marrow cells using the method of Dobson et al. (21) by placing 
the bones in 1.5-mL microfuge tubes supported by plastic inserts fabricated from 
either 0.5 mL microfuge tubes or hypodermic needle casings (Fig. 2). (see Note 5). 

4. Centrifuge the tubes briefly at 900 g for 5 sec in a microfuge. The marrow cells 
will pellet at the bottom of the tube. 

5. Resuspend the cell pellet in 1 mL of medium and create a single-cell suspension 
by aspirating through a 21 -gauge needle. 



2/26/03, 10:44 AM 

I ■ I I I 


Mineralizing Fibroblast-Colony-Forming Assays 33 

6. Pool the cells from all bones and make up to 10-20 mL with medium (see Note 6). 

7. Check the cell concentration using a hemocytometer (see Note 7). 

3.2. Setting up the CFU-f assay 

1. Plate 2 x 10 6 bone marrow cells out in 55-cm 2 Petri dishes in 10 mL of culture 
medium containing the required concentrations of the substance(s) to be tested. 

2. Change the medium after 5 d and thereafter twice weekly for up to 18 d. 

3. Terminate the cultures by washing with PBS and fix by adding cold 100% ethanol. 

3.3. ALP Staining 

1 . Add enough ALP staining solution to cover the Petri dish (~5 mL) and allow to 
stand for 30 min at room temperature. 

2. Wash the Petri dish under running tap water, and allow to dry. 

3. Photograph the dishes to document the amount of ALP staining. 

4. Destain the plates by gently shaking (approx 30 rpm) overnight with 100% etha- 
nol on an orbital shaker. 

3.4. Mineralization 

1 . Add enough alizarin red solution to cover each Petri dish (~5 mL) and allow to 
stand for 30 min at room temperature with gentle agitation. 

2. Wash the Petri dishes under running tap water until the excess dye has washed off. 

3. Allow to dry at room temperature. 

4. Photograph the dishes to document the amount of mineralization. 

5. Destain the dishes by gently shaking with 5% percholoric acid for 5 min. 

6. Wash thoroughly with tap water. 

3.5. Collagen 

1. Add 5 mL of collagen staining solution to each dish and incubate for 18 h. 

2. Wash the dishes under running tap water until the excess dye has washed off. 

3. Allow to air-dry at room temperature. 

4. Photograph the dishes to document the amount of staining. 

5. Destain the dishes by gently shaking with collagen destaining solution. 

6. Wash thoroughly with tap water. 

3.6. Total Colonies 

1 . Wash the plates with borate buffer. 

2. Cover the plates with borate buffer containing 1% methylene blue for 30 min. 

3. Wash three times with borate buffer. 

4. Allow the cultures air-dry at room temperature. 

5. Photograph the dishes to document the numbers of methylene blue positive CFU-f. 

3.7. Methods of Quantitating Colony Number and Size 

Colony number and size distribution are best determined by image analysis 
(22). It should be noted that the method described here is just one of a number 

03/Scutt/29-40/F1 33 1 2/26/03, 10:44 AM 



Scutt et al. 


Fig. 3. A typical setup for the analysis of fibroblastic colonies. The stained cultures 
are photographed over a white-light transilluminator using a Kodak DC50 digital cam- 
era and the images downloaded to a computer. The images are then processed and 

of image analysis based methods of colony counting (23-26) all of which suf- 
fer from the weakness of being unable to resolve colonies at the periphery of 
the culture vessel and colonies that merge into each other. Automated methods 
perform considerably better than manual ones in that they are totally objective 
and do not suffer from operator fatigue. The system described here was chosen 
because much of the hardware is already in use in most laboratories and the 
software is freely available and relatively cheap. 

3.8. Acquiring Images for Colony Quantitation 

1 . Place the Petri dishes in an appropriately sized template over a white light trans- 

2. Acquire the image using a Kodak DC50 digital camera mounted on a camera 
(Fig. 3; see Note 8). 

3.9. Processing the Image 

1 . To analyze the images obtained with the digital camera (Fig. 4A,E) they first have 
to be converted to a format compatible with the IQ software using Photoshop 4.0. 

2. Owing to the uneven surface of the individual colonies, the IQ software will 
recognize larger colonies as several distinct colonies. This is rectified by apply- 
ing to the image a Gaussian blur with a radius of two pixels (Fig. 4B,F) and then 
a median filter with a radius of two pixels (Fig. 4C,G). This has the effect of 



2/26/03, 10:44 AM 

I ■ I I I 

Mineralizing Fibroblast-Colony-Forming Assays 



Fig. 4. Processing the acquired image. In the unprocessed images (A, E), the sur- 
face of the individual colonies is somewhat uneven and would be recognised as sev- 
eral distinct colonies. This is rectified by applying a gaussian blur with a radius of two 
pixels (B, F) and then a median filter with a radius of two pixels (C, G). This removes 
small variations in intensity and isolated pixels whose values differ from those of their 
surroundings. To make the analysis quantitative, the image levels are adjusted such 
that the background intensity is set to zero and black is set to 255 (D, H). The image is 
then converted to an 8-bit grayscale TIFF image, saved, and analyzed. 

removing small variations in intensity by evening out differences between adja- 
cent pixels and also removing isolated pixels whose values differ from those of 
their surroundings. To make the analysis quantitative, adjust the image levels 
such that the background intensity is set to zero, and black (i.e., the area where 
the Petri dish is masked off) is set to 255 (Fig. 4D,H). 
3. Lastly, convert the image to an 8-bit grayscale TIFF image and save it. The IQ 
software, like many image analysis packages, can analyze only grayscale images. 

3. 10. Analyzing the Image Using the IQ Program 

Import the 8-bit grayscale images into the IQ program (see Note 9). Mark 
the area of interest and analyze this using the colony counting mode. Colonies 
can be selected according to both their size and intensity. In the study shown in 
Fig. 4, colonies of at least 20 pixels (corresponding to 1 mm) in diameter and 
having an intensity of at least 20 gray levels above background (corresponding 
to approx 80 cells) were selected. The software then assigns an identity to each 
colony and calculates its coordinates, surface area, and intensity (Table 1). 

4. Notes 

1. Choice of medium: Other media can be used instead of DMEM, the main crite- 
rion being the presence of 4500 mg/L of glucose. oc-MEM can also be used, 
although in pharmacological studies it is difficult to get an increase in colony 
number using bone anabolic drugs with this medium. It should be noted that the 
absence of phenol red leads to decreased colony number and differentiation. The 
reason for this is unknown and does not appear to be related to its estrogenic 



2/26/03, 10:44 AM 

I ■ I I I 


Scutt et al. 

Table 1 

Extract from the Report Generated After Analysis of the Petri Dish 

Shown in Fig. 3 


Bio Image Colony 

Data Report 





Time: 11:56:25 

Number of colonies 

: 130 

Area: 4143.71 

% Area: 8.96 

Colony # Colony x 


y Int. Bkgd. 


% Area Add Colony Name 




0.51 0.02 







0.05 0.04 







0.30 0.04 







0.18 0.02 







0.27 0.04 







0.09 0.03 







0.14 0.02 







0.39 0.02 







0.35 0.01 







0.58 0.01 







0.33 0.01 







0.05 0.03 







0.05 0.03 







0.43 0.02 







0.33 0.01 




The data for the first 15 colonies is shown and it can be seen that a total of 130 colonies were 
detected and that these covered 8.96% of the area selected. The colony coordinates, intensity, and 
area are given for each colony and these data can be imported into packages such as Excel and 
analyzed statistically. 

properties. For reasons that are unclear, some batches of FCS will not support 
collagen accumulation or calcification. This be rectified by using 5% FCS and 
5% "serum supreme," a supplemented newborn calf serum. Some laboratories 
use ascorbate-2-phosphate in place of ascorbate but we have found that this leads 
to a nonspecific overcalcification. 

2. Fixation: Industrial methylated spirit can be used instead of 70% ethanol. 

3. Other methods of ALP staining: Other substrates and dyes can be used for ALP 
staining such as naphthol phosphate AS-MX, fast red TR, or fast blue B, but it is 
not possible to destain them adequately. 

4. Methods of euthanasia and sex of animals: Other methods of euthanasia may 
also be used. It is best to use male rats, as the hormonal variations during the 
estrous cycle in female rats can have a pronounced effect on the colony number 
achieved. Surgery (e.g., orchidectomy, ovariectomy, and sham procedures) have 
dramatic effects on colony number that last several weeks. 



2/26/03, 10:44 AM 

I ■ I I I 


Mineralizing Fibroblast-Colony-Forming Assays 37 

5. Methods of harvesting bone marrow cells: The centrifugation method is opti- 
mized for use with rat bones; however, with very minor modifications it can be 
adapted for use with mouse bones and with human bone chips. Another method 
of harvesting is to flush the cells from the bone using a syringe but the variability 
of the yield will be greatly increased. This is of particular importance when per- 
forming ex vivo studies, in which the near 100% recovery of the centrifugation 
method improves reproducibility greatly. 

6. Volumes of culture medium used to harvest cells: When carrying out ex vivo cul- 
tures, changes in the numbers of other bone marrow cell types can have a con- 
founding effect on the frequency of CFU-f. For this reason, when carrying out 
these kinds of experiments the cells are recovered in a constant volume of culture 
medium (10 mL per bone for a 200-g rat) and 0.5 mL of the resultant suspension is 
used in the culture. In this way the number of CFU-f per bone can be calculated and 
is largely independent of changes in the overall bone marrow cell population. 

7. The overall cell viability is between 70% and 80%. 

8. Acquiring images for colony quantitation: Position the camera stand so that the 
Petri dish fills as much of the image as possible. With careful positioning, images 
from dishes that have been stained sequentially can be easily aligned. There are 
now a good selection of digital cameras suitable for this type of work. The main 
criteria are the size of the memory and the speed of downloading. We have found 
that increasing the resolution beyond 800 x 600 pixels is unnecessary. 

9. Image analysis programs: The IQ Bioimage software program described here is 
no longer available. There are a number of other cheap packages available dedi- 
cated to colony counting such as MACE (Weiss Associates, Branford, CT, USA; and a system is also being developed by The Gray Labo- 
ratory (Middlesex, UK; see ref. 26) which will soon be on the market. Alterna- 
tively, with some ingenuity, other software can be adapted to count colonies such 
as NIH image or the Leica Q Win software. There are also a number of dedicated 
colony counting systems on the market that, although designed to count bacterial 
colonies, will also count fibroblastic colonies. They are somewhat expensive, 
being of the order of £10,000 (approx $15,000) or more. 


1. Mundy, G. R. (1995) No bones about fluoride. Nat. Med. 1, 1130-1131. 

2. Friedenstein, A. J. (1990) Osteogenic stem cells in the bone marrow. Bone Miner. 
Res. 7, 243-272. 

3. Maniatopoulos, C, Sodek, J., and Melcher, A. H. (1988) Bone formation in vitro 
by stromal cells obtained form bone marrow of young rats. Cell Tissue Res. 254, 

4. Nefussi, J-R., Boy-Lefevre, M. L., Boolekbache, H., and Forest, N. (1985) Miner- 
alization in vitro of matrix formed by osteoblasts isolated by collagenase diges- 
tion. Differentiation 29, 160-168. 

5. Bellows, C. G., Aubin, J. E., Heersche, J. N. M. and Antosz, M. E. (1986) 
Mineralised bone nodules formed in vitro from enzymatically released rat calva- 
rial cell populations. Calcif. Tissue Int. 38, 143-154. 

03/Scutt/29-40/F1 37 I 2/26/03, 10:44 AM 

I ■ I I I 


38 Scutt et al. 

6. Stein, G. S., Lian, J. B. and Owen, T. A. (1990) Relationship of cell growth to the 
regulation of tissue specific gene expression during osteoblast differentiation. 
FASEBJ. 4, 3111-3123. 

7. Bellows, C. G. and Aubin, J. E. (1989) Determination of numbers of 
osteoprogenitors present in isolated fetal rat calvarial cells in vitro. Dev. Biol. 
133, 8-13. 

8. Scutt, A. and Bertram, P. (1995) Bone marrow cells are targets for the anabolic 
actions of prostaglandin E 2 on bone: induction of a transition from non-adherent 
to adherent osteoblast precursors. /. Bone Miner. Res. 10, 474-489. 

9. Still, K. and Scutt, A. (2001) Stimulation of CFU-f formation by prostaglandin E 2 
is mediated in part by its degradation product, prostaglandin A 2 . Prostaglandins 

10. Pitaru, S., Kotov-Emeth, S., Noff, D., Kaffuler, S., and Savion, N. (1993) Effect 
of basic fibroblast growth factor on the growth and differentiation of adult stro- 
mal bone marrow cells: enhanced development of mineralized bone-like tissue in 
culture. /. Bone Miner. Res. 8, 919-929. 

11. Gronthos, S., Simmons, P. J., Graves, S. E., and Robey, P. G. (2001) Integrin- 
mediated interactions between human bone marrow stromal precursor cells and 
the extracellular matrix. Bone 28, 174-81. 

12. Walsh, S., Jordan, G. R., Jefferiss, C, Stewart, K., and Beresford, J. N. (2001) 
High concentrations of dexamethasone suppress the proliferation but not the dif- 
ferentiation or further maturation of human osteoblast precursors in vitro: rel- 
evance to glucocorticoid-induced osteoporosis. Rheumatology 40, 74-83. 

13. Scutt, A., Kollenkirchen, U., and Bertram P. (1996) The effect of age and ovariec- 
tomy on fibroblastic colony-forming unit numbers in rat bone marrow. Calcif. 
Tissue Int. 59, 309-310. 

14. Erben, R. G., Scutt, A. M., Miao, D., Kollenkirchen, U., and Haberey, M. (1997) 
Short-term treatment of rats with high-dose calcitriol stimulates bone formation 
in vivo and increases the number of osteoblast precursor cells in the bone marrow. 
Endocrinology 138, 4629-4635. 

15. Nishida, S., Yamaguchi, A., Tanizawa, T., et al. (1994) Increased bone formation 
by intermittent parathyroid hormone administration is due to the stimulation of 
proiliferation and differentiation of osteoprogenitor cells in the bone marrow. 
Bone 15, 717-723. 

16. Weinreb, M., Suponitzky, I., and Keila, S. (1997) Systemic administration of an 
anabolic dose of PGE2 in young rats increases the osteogenic capacity of bone 
marrow. Bone 20, 521-526. 

17. Jilka, R. L., Weinstein, R. S., Takahashi, K., Parfitt, A. M., and Manolagas, S. C. 
(1996) Linkage of decreased bone mass with impaired osteoblastogenesis in a 
murine model of accelerated senescence. /. Clin. Invest. 97, 1732-1740. 

18. Di Gregorio, G. B., Yamamoto. M., Ali, A. A., et al. (2001) Attenuation of the 
self -renewal of transit-amplifying osteoblast progenitors in the murine bone mar- 
row by 17 beta-estradiol. /. Clin. Invest. 107, 803-812. 

03/Scutt/29-40/F1 38 1 2/26/03, 10:44 AM 

I ■ I I I 


Mineralizing Fibroblast-Colony-Forming Assays 39 

19. Jilka, R. L., Takahashi, K., Munshi, M., Williams, D. C, Roberson, P. K., and 
Manolagas S. C. (1998) Loss of estrogen upregulates osteoblastogenesis in the 
murine bone marrow. Evidence for autonomy from factors released during bone 
resorption. /. Clin. Invest. 101, 1942-1950. 

20. Kajkenova, O., Lecka-Czernik, B., Gubrij, I., et al. (1997) Increased adipogenesis 
and myelopoiesis in the bone marrow of SAMP6, a murine model of defective 
osteoblastogenesis and low turnover osteopenia. /. Bone Miner. Res. 12, 1772-1779. 

21. Dobson, K. R., Reading, L., Haberey, M., Marine, X., and Scutt, A. (1999) Cen- 
trifugal isolation of bone marrow from bone: an improved method for the recov- 
ery and quantitation of bone marrow osteoprogenitor cells from rat tibiae and 
femurae. Calcif. Tissue Int. 65, 41 1-413. 

22. Dobson, K., Reading, L., and Scutt, A. (1999) A cost effective method for the 
automatic quantitative analysis of fibroblastic-colony forming units with osteo- 
blastic potential. Calcif. Tissue Int. 65, 166-172. 

23. Parry, R. L., Chin, T. W., and Donahoe, K. (1991) Computer-aided cell colony 
counting. BioTechniques 10, 112-11 A. 

24. Nefussi, J. R., Ollivier, A., Oboeuf, M., and Forest, N. (1997) Rapid nodule evalu- 
ation computer-aided image analysis procedure for bone nodule quantification. 
Bone 20, 5-16. 

25. Hoekstra, S. J., Tarka, D. K., Kringle, R. O., and Hincks, J. R. (1998) Develop- 
ment of an automated bone marrow colony counting system. In Vitro Mol. Toxicol. 
11, 207-213. 

26. Barber, P. R., Vojnovic, B., Kelly, J., et al. (2001) Automated counting of mam- 
malian colonies. Phys. Med. Biol. 46, 63-76. 

03/Scutt/29-40/F1 39 1 2/26/03, 10:44 AM 

I ■ I I I 




2/26/03, 10:44 AM 

I ■ I I I 


Osteocyte Isolation and Culture 

Peter J. Nijweide, Arie van der Plas, Marcel J. Alblas, 
and Jenneke Klein-Nulend 

1. Introduction 

Osteocytes are the most abundant cells in bone. Although individual osteo- 
cytes are buried in an isolated position within bone matrix, they remain in con- 
tact with one another and with cells on the bone surface by long cell processes 
that run via small channels, termed canaliculi, through the bone matrix. Where 
the cell processes of two osteocytes meet in a shared canaliculus, gap junctions 
provide intracellular contact (1). For a long time osteocytes were outside the 
mainstream of bone research. Increasing interest in the mechanoregulation of 
bone has changed this, and today there is a general consensus that osteocytes 
play a pivotal role as mechanosensors and effectors in bone (2). Whether 
osteocytes have other functions remains to be elucidated. 

The anatomical location of osteocytes deep within bone has proved to be a 
major obstacle in studying the role that osteocytes play in bone metabolism. 
Osteocytes depend for their activities and survival on the diffusion of oxygen, 
hormones, nutrients, and waste via the canaliculi. The bone tissue culture meth- 
ods that have been used so successfully in the studies of osteoblast and osteo- 
clast activity are therefore generally of little value in the evaluation of osteocyte 
function. Bone explants have to be very limited in size to allow sufficient trans- 
port of nutrients to and from the osteocytes, and in fetal bone tissues the osteo- 
cyte contribution to the cellular component is very small. 

Direct isolation of osteocytes is therefore the method of choice to study 
osteocyte physiology (3). Three major problems arise, however. First, how does 
one isolate live osteocytes from the bone matrix in sufficient numbers to study? 
Second, how can osteocytes be separated from other cells and how can the 
population be kept homogeneous? Third, how can osteocytes be recognized in 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


04/Nijweide/41-50/F1 41 I 2/26/03, 10:45 AM 

I ■ I I I 


42 Nijweide et al. 

culture, as the cells tend to lose some of their morphological characteristics 
when they are removed from their three-dimensional tissue structure? 

This chapter describes the isolation of osteocytes from 18-d-old fetal chicken 
calvariae. Calvariae of this age are used because at that developmental stage 
the calvariae are not yet too heavily calcified to allow collagenase to liberate 
matrix-entrapped cells. The isolation of these cells is facilitated by mild EDTA 
treatments alternating with collagenase treatments. The choice of fetal chicken 
calvariae is important because in this species the periostea from both calvaria 
surfaces can be dissected relatively easily, which results in the absence of most 
periosteal cells in the collagenase released cell population. Furthermore, fetal 
chicken calvariae have a much broader layer of osteoid on their surfaces than 
mouse or rat calvariae. Finally, the availability of a monoclonal antibody 
(MAb) that specifically recognizes osteocytes allows the purification of osteo- 
cytes from mixed cell populations (immunodissection). The MAb used in the 
separation procedure is also used to recognize osteocytes in cell cultures. 

2. Materials 

2. 1. Fertilized Chicken Eggs 

Incubate the eggs at 38.5°C in a humid air atmosphere for 18 d. The incuba- 
tor should have a mechanism that turns the eggs regularly through 180°, about 
10 times per hr. Fertilized eggs can be stored for 2-3 wk at 14-16°C before 
development of the embryo is started by incubation in the incubator. 

2.2. Media and Solutions 

1. Hanks' balanced salt solution (HBSS) from Gibco-BRL Life Technologies. 

2. Phosphate-buffered salt solution (PBS): 137 mM NaCl, 2.7 mM KC1, 8.1 mM 
Na 2 HP0 4 and 1.5 mM KH 2 P0 4 . Adjust to pH 7.4. 

3. Isolation salt solution (ISS)(4): 70 mM NaCl, 30 mM KC1, 1 mM CaCl 2 , 10 mM 
NaHC0 3 , 25 mM W-2-hydroxyethylpiperazine-W-2-ethanesulfonic acid 
(HEPES), 5 mg/mL of glucose (Sigma), and 1 mg/mL of bovine serum albumin 
(BSA) (ICN Biomedicals Inc.). Adjust to pH 7.4 at 37°C. 

4. Isolation medium: Add 7 u,moles/L of W a -tosyl-L-lysyl-chloromethane hydrochlo- 
ride (BDH Biochemicals) and 1 mg/mL of collagenase Type I (Sigma) to ISS. 
The Af a -tosyl-L-lysyl-chloromethane hydrochloride is added to inhibit proteases 
other than collagenase (4). 

5. EDTA solution: 4 mM EDTA in PBS. Adjust pH to 7.4. 

6. Wash fluid: 10% Inactivated chicken serum (Sigma) in a-minimum essential 
medium (oc-MEM) (Gibco). For inactivation, heat serum for 30 min at 56°C and 
centrifuge at 200g for 5 min. 

7. Culture medium: a-MEM fortified with 2% inactivated chicken serum, 0.2 g/L 
of glutamine (Sigma), 0.05 g/L of ascorbic acid (Sigma), 0.05 g/L of gentamicin 
(Sigma) and 1 g/L of glucose. 

04/Nijweide/41-50/F1 42 I 2/26/03, 10:45 AM 

I ■ I I I 


Osteocyte Isolation 43 

8. Trypsin-EDTA (TE) solution: 0.05% Trypsin 1 :250 (Sigma) and 0.27 mM EDTA 
in PBS. 

9. Coated bead suspension: Mix DNA-conjugated beads (CELLection™ Pan Mouse 
IgG kit from DYNAL), with MAb OB7.3 IgG (see Subheading 2.3.) and PBS to 
a final concentration of 15 \ig of IgG / 8 x 10 7 beads/mL. Two hundred and fifty 
microliters of this suspension is needed for the isolation of osteocytes from 40 
calvariae. Incubate overnight at 4°C, gently shaking the suspension, and store at 
4°C until use. Wash the beads shortly before use two times with 2% chicken 
serum in HBSS. For this the bead suspension is put in a holder next to a magnet 
(DYNAL) that attracts the magnetic beads to one side of the tube, allowing the 
removal of the suspension fluid. Finally, resuspend the IgG-coated beads in 250 u.L 
of 2% chicken serum in HBSS. 

10. Zinc fixative: Dissolve 0.5% zinc chloride and 0.5% zinc acetate in 0.1 M Tris- 
acetate buffer, pH 4.5. 

2.3. MAb OB7.3 

The antibody was originally raised according to standard procedures by 
injecting bone cells isolated from calvariae of 18-d-old fetal chickens into 
BALB/c mice (5). In bone sections it recognizes only osteocytes embedded in 
osteoid or in calcified matrix (Fig. 1A,B). In cultures of cells enzymatically 
isolated from fetal chicken calvariae MAb OB7.3 stains a minority of the iso- 
lated cells. The positive cells show an osteocyte-like morphology in culture 
(Fig. 1C,D and Fig. 2). The antibody may be obtained from either Dr. J. Klein- 
Nulend (Deptartment of Oral Cell Biology, ACTA-Vrije Universiteit, 
Amsterdam, The Netherlands) or from Dr. K. E. de Rooij (Deptartment of 
Endocrinology and Metabolism, Leiden University Medical Center, Leiden, 
The Netherlands). 

3. Methods 

3. 1. Tissue Dissection 

1. Remove the eggs from the incubator after 18 d. Keep one of the eggs with the 
blunt side upwards where the air chamber is located. Crack the top of the shell 
and peel the shell off to the edge of the air chamber with a pair of sterile tweezers . 
Stab one leg of a second pair of tweezers underneath the white shell membrane 
and the chorioallantoic membrane. Close the tweezers and peel both membranes 
off in one movement. 

2. Grasp the embryo under its head with a curved forceps, lift it a little above the egg, 
and decapitate underneath the forceps. The body of the embryo will fall back into 
the egg. Transfer the head to a Petri dish with some HBSS placed on ice. Resterilize 
tweezers and forceps by dipping into ethanol (100%) and burning the ethanol off. 

3. Make a cut in the back of the neck with a small pair of scissors. Hold the head by 
the bill with tweezers and tear the skin off in the direction of the bill. Cut the 
calvaria loose along the edge and cut it into two parts through the central suture. 

04/Nijweide/41-50/F1 43 1 2/26/03, 10:45 AM 

I ■ I I I 


Nijweide et al. 


Fig. 1. Frozen section of a calvaria (A, B) and air-dried cell cultures (C-F) stained 
with MAb OB7.3. Left side: Phase contrast. Right side: Immunofluorescence. (A,B) 
Section of an 18-d-old fetal chicken calvaria. Note that only the osteocytes inside the 
bone matrix are stained. (C,D) OBmix after 1-d of culture. In the lower left corner a 
group of osteocytes with one MAb OB7.3-negative cell {white asterisk); in the upper 
right corner an osteoblast colony with two osteocytes. (E,F) Purified osteocyte popu- 
lation. Note the contaminating fibroblast-like cell in the upper left corner (white aster- 
isk). Scale bar = 100 ^m. (Reproduced from /. Bone Miner. Res. 1992; 7, 389-396 
with permission of the American Society for Bone and Mineral Research.) 

4. Put the calvaria halves in a droplet of sterile HBSS and remove the ectocranial 
and endocranial periostea under a dissecting microscope with small scalpels. 


2/26/03, 10:45 AM 

I ■ I I I 

Osteocyte Isolation 



Fig. 2. Scanning electron micrographs of isolated osteocytes. (A) Osteocyte 5 min 
after seeding. The osteocyte is already attached and has formed finger-like projections 
in all directions. (B) Osteocytes 20 min after seeding. The cell processes in the plane 
of the support have elongated, whereas the processes perpendicular to the support 
have disappeared. Note the presence of two magnetic beads. (C) Two osteocytes have 
made contact with each other via their cell processes after 24 h of culture. (D) Exten- 
sive network of flattened osteocytes with many branched cell processes after 48 h of 
culture. Scale bar = 10 \im. (Reproduced from /. Bone Miner. Res. 1992; 7, 389-396 
with permission of the American Society for Bone and Mineral Research.) 

Transfer the "bare" calvaria halves into a Petri dish with HBSS and keep on ice. 
5. Repeat these actions for all eggs. Generally we use 40 eggs in one isolation ses- 
sion. Too many eggs would make the dissection procedure too long; a too small 
number of eggs is inefficient and relatively decreases the number of isolated 

3.2. Isolation of OBmix 

Perform all incubations in the steps below in a shaking water bath set at 37°C. 

1 . Transfer all calvaria halves to a small flask containing 3 mL of isolation medium 
diluted 10 times with ISS (final collagenase concentration: 0.1 mg/mL). Incubate 
for 10 min and discard the supernatant. Repeat this step once. (The discarded 
supernatants will contain primarily damaged cells and erythrocytes.) 



2/26/03, 10:45 AM 

I ■ I I I 


46 Nijweide et al. 

2. Add 3 mL of isolation medium (final collagenase concentration: 1 mg/mL) to the 
calvariae. Incubate for 15 min and discard the supernatant. (This will contain 
primarily fibroblastic cells and some osteoblasts.) Wash calvariae three times 
with 2.5 mL of PBS each time; discard the PBS washings. 

3. Add 4 mL of EDTA solution to the calvariae. Incubate for 10 min and remove the 

4. Centrifuge at 4°C for 3 min at 200g and resuspend the cell pellet in wash fluid. 
This cell suspension is termed fraction 1. 

5. Wash the calvariae three times, once with 2 mL of PBS, and twice with 1 mL of 
ISS. Add the washings to fraction 1 from step 4 and keep on ice. 

6. Add 4 mL of isolation medium to the calvariae and incubate for 45 min. Remove 
the supernatant; centrifuge at 4°C for 3 min at 200g. 

7. Resuspend the sedimented cells in wash fluid. This cell suspension is termed 
fraction 2. 

8. Wash the calvariae three times with 1 mL of PBS. Add the washings to fraction 2 
from step 7. 

9. Combine fractions 1 and 2, centrifuge at 4°C for 3 min at 200g and resuspend the 
sedimented cells in culture medium. This is the OBmix population. It contains 
osteoblasts, 20-30% osteocytes, and a few fibroblasts (Fig. 1C,D). 

10. Determine the cell concentration with, for example, a Burker-Tiirk hemocytom- 
eter and increase the volume of the cell suspension until a concentration of about 
2 x 10 6 cells/mL is obtained. 

1 1 . Seed 0.5 mL of the cell suspension into small (50-mL) culture flasks containing 3 mL 
of culture medium per flask. Generally about four culture flasks (4 x 10 6 cells) 
will be needed. 

12. Culture the cells for 24 h at 37°C in a humid atmosphere of 5% C02 in air (see 
Note 1). 

3.3. Isolation of Osteocytes 

1. Remove the culture medium from the culture flasks on the next day, and rinse 
three times with PBS. Add 3 mL of TE solution per flask and incubate for 3 min 
at 37°C. Stop the trypsin activity by adding 0.3 mL of chicken serum per flask 
and hit the lab table three times with the flasks to loosen the OBmix cells from 
the bottom. 

2. Disperse the cells by repeated pipetting over the surface bearing the cell layer. 

3. Remove cell clumps by sieving the cell suspension through a nylon sieve of 30-jJ.m 
mesh, sediment the cells at 200g for 3 min, discard the supernatant and resuspend 
the cells in 2 mL of cold (4°C) 2% chicken serum in HBSS. The sieving proce- 
dure is necessary to achieve a single cell suspension. OBmix cells, in particular 
the osteocytes, tend to form clumps of (probably) gap junction-coupled cells. 
These clumps will, however, contain not only osteocytes but also osteoblasts that 
will later quickly overgrow the nondividing osteocytes (see Note 2). Centrifuge 
the cell suspension (200g, 3 min) and resuspend the sedimented cells in 125 [ih 
of HBSS-2% serum. 

04/Nijweide/41-50/F1 46 1 2/26/03, 10:45 AM 

I ■ I I I 


Osteocyte Isolation 47 

4. Add 125 u,L of coated bead suspension, and incubate for 15 min in a rotator (60 
rpm) at 4°C. Take good care during this step that beads and cells remain in sus- 

5. Separate the bead-bound osteocytes from the osteoblasts with the magnet. Wash 
the bead-bound osteocytes four times with 2% chicken serum in HBSS (using the 
magnet to collect the osteocytes) and finally resuspend the osteocytes in 200 uE 
of 2% chicken serum in HBSS. 

6. Add a fresh batch of 125 uE of coated beads to the osteoblast fraction and repeat 
the separation procedure. 

7. Combine both two bead-bound osteocyte fractions, separate the cells with the 
magnet, and resuspend them in 100 u.L of culture medium. Count the number of 
cells per unit volume. The bead-bound osteocytes can now either be seeded and 
cultured or can be freed from their beads and cultured (see Note 3). 

8. For immediate removal of the beads, wash the bead-bound osteocytes once with 
PBS-2% chicken serum, separate the cells with the magnet, and resuspend them 
in 100 uE of PBS-2% chicken serum. 

9. Add 4 uL of releasing buffer containing 50 U of DNase/uE (CELLection™ Pan 
Mouse IgG kit) and incubate 15 min at 37°C in a shaking water bath. 

10. Separate the beads from the osteocytes with the magnet. Wash the beads two 
times with PBS-2% chicken serum to remove all osteocytes from the beads and 
add the washings to the liberated osteocytes. Remove the last contaminating beads 
with the magnet. 

11. Centrifuge the cell suspension, discard the supernatant, and resuspend the osteo- 
cyte pellet in culture medium. 

12. If it is not necessary to use the isolated osteocytes immediately, it is much easier 
to seed the bead-bound osteocytes in culture medium in a Petri dish. The next 
day, the beads can simply be removed by washing the cell layer. DNase treat- 
ment is not needed. The osteocyte yield from 40 calvariae is generally approx 
200,000 osteocytes (Fig. 1E,F). 

13. Attached osteocytes can be removed from their support by a short treatment with 
TE solution. After washing and reseeding, the osteocytes reacquire their typical 
morphology of stellate cells in secondary culture (see Note 4). 

4. Notes 

1. Optimizing isolation of OBmix: The actual procedure one has to use depends 
heavily on the activity of the collagenase. If you start with a new batch of colla- 
genase, you may have to adapt the procedure. For example, step 2 may have to be 
shortened, if too many osteocytes are isolated in this step. If too many fibroblasts 
are still present in fraction 1, step 2 should be repeated. Steps 3-6 have to be 
repeated if many cells, in particular osteocytes, are isolated in these repeated 
steps. The yield of OBmix cells from 40 calvariae following the protocol in Sub- 
heading 3.2. is generally 3-4 x 10 6 cells. About 20-30 % of these cells are osteo- 
cytes (i.e., about 10 6 osteocytes). Most of the OBmix cells are, however, present 
in clumps of osteoblasts, osteocytes, or mixtures of osteoblasts and osteocytes. 

04/Nijweide/41-50/F1 47 I 2/26/03, 10:45 AM 

I ■ I I I 


48 Nijweide et al. 

Trying to isolate osteocytes from this population results primarily in the isolation 
of clumps of osteocytes often containing osteoblastic cells with high prolifera- 
tive capacity. Sieving the original OBmix population (see Subheading 3.2., step 
9) before immuno-isolation of osteocytes results in a very low osteocyte yield. 
Culturing the OBmix population for 24 h allows the cells of the clumps to sepa- 
rate themselves from each other, increases the number of single cells after the 
trypsin-EDTA treatment and therefore increases the osteocyte yield. 

2. Reducing numbers of contaminating cells: Generally, we use 2% or less chicken 
serum in the culture medium. Osteocytes are postmitotic cells. Because serum 
strongly stimulates cells capable of mitosis to proliferate, isolated osteocytes will 
therefore be quickly overgrown in the presence of serum by contaminating cells. 
A too low serum concentration will, however, deteriorate cell quality. In view of 
this, we recommend that osteocyte cultures be used for experimentation as soon 
as possible after isolation. 

3. Purity of the osteocyte population. If the isolation procedure is used routinely, it 
is advisable to examine regularly the purity of the osteocyte populations. Con- 
tamination of the osteocyte isolate may be determined by seeding a small sample 
of the osteocyte suspension in a culture dish. Incubate until the cells are firmly 
attached (4-6 h), and immunostain the cells for the presence of MAb 7.3. Note 
that removal of the beads will not remove the antibody from the cell surface! At 
least 95% of the cells should be positive for the antibody (Fig. 1E,F). 

4. Phenotypic characterization of osteocytes: In bone, osteocytes are fully defined 
by their location within the bone matrix. For isolated osteocytes other markers 
are needed to establish their identity. 

a. Stellate morphology. Avian osteocytes reacquire the stellate morphology of 
osteocytes in situ after isolation and attachment (6) (Fig. 2). This may be also 
the case for murine osteocytes (7). 

b. Antibodies. Three osteocyte-specific MAbs have been described in the litera- 
ture: MAb OB7.3 (5), MAb OB37.11 (8), and MAb SB5 (9). All three are 
specific for avian osteocytes and do not cross-react with mammalian cells. 
The identities of the three antigens involved have not been reported, although 
that of MAb OB7.3 is recently elucidated (10). So far, MAb OB7.3 is the only 
antibody used for osteocyte isolation. Other antibodies, but apparently less 
specific, are El 1, a monoclonal antibody that reacts specifically with highly 
mature osteoblasts and with osteocytes in tissue sections of rat bone (11) and 
antibodies against fimbrin, an actin-bundling protein that is abundantly 
present in osteocytes (12). The OB7.3 antigen is easily destroyed by harsh 
fixation procedures. For the immunostaining of osteocytes in tissues unfixed, 
air-dried, frozen sections are recommended (Fig. 1A,B). Cell cultures can be 
stained by incubating live cells with MAb OB7.3 in HBSS for 30-60 min at 
room temperature or by washing the cells with HBSS followed by air drying 
and incubation with the antibody (Fig. 1C-F). Also a short (10 min) fixation 
with 2-4% paraformaldehyde at 4°C will leave enough antigen intact for a 
reasonable staining. More recently, we have used zinc fixative successfully 

04/Nijweide/41-50/F1 48 1 2/26/03, 10:45 AM 

I ■ I I I 


Osteocyte Isolation 49 

for the fixation and immunostaining of cell cultures. Zinc fixative does not 
appear to damage the OB7.3 antigen. A further advantage of this fixative is 
that cell cultures can be kept in the fixative for a long time (days) without loss 
of immunoreactivity of the antigen. An osteocyte specific antibody similar to 
MAb OB7.3 but reacting with nonavian cells is as yet not reported. Mikuni- 
Takagaki et al. have used a repeated collagenase treatments of newborn rat 
calvariae. They describe that their last, seventh fraction contains many cells 
with an osteocyte-like morphology, 
c. Protein products. Several proteins, such as osteocalcin and osteopontin, have been 
demonstrated in or around osteocytes in relatively high amounts (13). Alkaline 
phosphatase, a cell surface bound enzyme, is generally low in osteocytes, in par- 
ticular compared to the amounts present on osteoblasts. CD44, a membrane-bound 
glycoprotein that is involved in cell attachment to matrix proteins, is generally 
highly expressed in osteocytes (14), but is also expressed by other cells in bone. 
The production and presence of these proteins is not specific for osteocytes but 
can be used as additional markers for osteocyte identification. 


1. Doty, S. B. (1981) Morphological evidence of gap junctions between bone cells. 
Calcif. Tissue Int. 33, 509-512. 

2. Nijweide, P. J., Burger, E. H., Klein-Nulend, J., and van der Plas, A. (1996) The 
osteocyte, in Principles of Bone Biology (Bilezikian, J. P., Raisz, L. G., and 
Rodan, G. A., eds.), Academic Press, San Diego, pp. 115-126. 

3. Van der Plas, A., Aarden, E. M., Feijen, J. H. M., et al. (1994) Characteristics and 
properties of osteocytes in culture. /. Bone Miner. Res. 9, 1697-1704. 

4. Hefley, T. J. (1987) Utilization of FPLC-purified bacterial collagenase for the 
isolation of cells from bone. /. Bone Miner. Res. 2, 505-516. 

5. Nijweide, P. J. and Mulder, R. J. P. (1986) Identification of osteocytes in osteo- 
blast-like cell cultures using a monoclonal antibody specifically directed against 
osteocytes. Histochemistry 84, 342-347. 

6. Van der Plas, A. and Nijweide, P. J. (1992) Isolation and purification of osteo- 
cytes. /. Bone Miner. Res. 7, 389-396. 

7. Mikuni-Takagaki, Y., Kakai, Y., Satoyoshi, M., et al. (1995) Matrix mineraliza- 
tion and the differentiation of osteocyte-like cells in culture. /. Bone Miner. Res. 

8. Nijweide, P. J., van der Plas, A., and Olthof, A. A. (1988) Osteoblastic differen- 
tiation, in Cell and Molecular Biology of Vertebrate Hard Tissues, Ciba Founda- 
tion Symposium 136 (Evered, D. and Harnett, S., eds.), John Wiley & Sons, 
Chichester, UK, pp. 61-77. 

9. Bruder, S. P. and Caplan, A. I. (1990) Terminal differentiation of osteogenic cells 
in the embryonic chick tibia is revealed by a monoclonal antibody against osteo- 
cytes. Bone 11, 189-198. 

10. Westbroek, I., De Rooij, K. E., and Nijweide, P. J. (2002) Osteocyte-specific 
monoclonal antibody MAb OB7.3 is directed against Phex protein. /. Bone Miner. 
Res. 17, 845-853. 

04/Nijweide/41-50/F1 49 I 2/26/03, 10:45 AM 

I ■ I I I 

50 Nijweide et al. 

11. Wetterwald, A., Hoffstetter, W., Cecchini, M. G., et al. (1996) Characterization 
and cloning of the El 1 antigen, a marker expressed by rat osteoblasts and osteo- 
cytes. Bone 18, 125-132. 

12. Tanaka-Kamioka, K., Kamioka, H., Ris, H., andLim, S. S. (1998) Osteocyte shape 
is dependent on actin filaments and osteocyte processes are unique actin-rich pro- 
jections. /. Bone Miner. Res. 13, 1555-1568. 

13. Aarden, E. M., Wassenaar, A. M., Alblas, M. J., and Nijweide, P. J. (1996) Immu- 
nocytochemical demonstration of extracellular matrix proteins in isolated osteo- 
cytes. Histochem. Cell Biol. 106, 495-501. 

14. Nakamura, H. and Ozawa, H. (1996) Immunolocalization of CD44 and the ERM 
family in bone cells of mouse tibiae. /. Bone Miner. Res. 11, 1715-1722. 


04/Nijweide/41-50/F1 50 1 2/26/03, 10:45 AM 

I ■ I I I 


Isolated Osteoclast Cultures 

Astrid Hoebertz and Timothy R. Arnett 

1. Introduction 

Osteoclasts are large, multinucleated cells formed by the fusion of hemato- 
poietic, mononuclear progenitors of the monocyte/macrophage lineage, and 
are the cells responsible for resorbing bone. Osteoclasts are usually few in num- 
ber relative to other cell types in bone and are difficult to isolate because they 
are contained in a hard tissue; in addition, they are at the end of their prolifera- 
tion and differentiation cycle, presenting problems for the creation of osteo- 
clast cell lines. However, with the development, almost 20 years ago, of in 
vitro resorption pit formation models, using isolated primary, mature osteo- 
clasts and mineralized bone or dentine matrix as a substrate (1,2), considerable 
progress was made in our understanding of osteoclast biology. Data from such 
short term cultures complements that obtained from bone organ culture resorp- 
tion models and long-term cultures of osteoclast forming hematopoietic stem 
cells derived from marrow or peripheral blood. 

Boyde, Jones, Chambers, and colleagues developed the disaggregated 
osteoclast resorption assay in 1984 (1,2). Variants of these assays were then 
widely adopted to study osteoclasts isolated from neonatal rat, rabbit, or chick 
long bones (4,17; see chapters by Collin-Osdoby et al., Chapter 6, and Coxon 
et al., Chapter 7). The method used in each case is extremely simple: osteo- 
clasts are relatively abundant in the bones of neonatal animals (reflecting the 
requirement for rapid growth modeling) and can be released mechanically by 
fragmenting the bones in a suitable medium. This can be achieved by either 
mincing the bones or scraping the exposed endosteal surfaces of longitudinally 
split bones, and the resulting cell suspension is settled onto bone or dentine 
discs and, after rinsing in saline to remove non-adherent cells, cultured for 
about 24 h. Under suitable conditions, osteoclasts can then excavate resorption 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


05/Hoebertz/51-64/F1 53 1 2/26/03, 10:45 AM 

I ■ I I I 

54 Hoebertz and Arnett 

lacunae. Although this model has several limitations in attempts to study the 
whole physiological cascade of bone resorption, it provides an excellent tool 
for detailed studies of the cellular mechanisms involved in the destruction of min- 
eralized bone matrix, especially since the application of confocal microscopy to 
study osteoclasts cultured on bone or dentine slices. Because osteoclasts sediment 
or adhere somewhat more rapidly than other cell types present in the mixed cell 
population released from fragmented bones, "functionally purified" osteoclast 
populations may be generated by careful adjustment of settling times and washing 
methods. Clearly, one of the most important factors in this assay system is to obtain 
adequate basal levels of resorption. This is accomplished by the use of slightly 
acidified culture medium, as first described by Amett and Dempster (4). 

Assessment of resorption is typically achieved by simply counting the number 
of multinuclear (more than three nuclei) osteoclasts, stained histochemically for 
tartrate-resistant acid phosphatase, and the number and/or area of resorption pits, 
using the technique of reflected light microscopy, after staining the discs with tolui- 
dine blue to visualize pits (5,17). This replaced the more complicated use of scan- 
ning electronic microscopy (SEM) to study resorption pits. Measuring the volume 
of each individual pit rather than discrete pit number or resorbed area is a more 
accurate method of assessing resorption (6) but is clearly more time consuming. 
-®- This chapter describes the isolation and short-term culture of osteoclasts 

obtained from neonatal rat long bones, which in principle resembles the isola- 
tion of osteoclasts from rabbit and chick long bones. A special focus is on the 
role of extracellular pH in osteoclast cultures (see Note 1). 

2. Materials 

1. Tissue culture medium: Minimum essential medium (MEM) supplemented with 
Earle's salts, 10% fetal bovine serum, 2 mM L-glutamine, 100 U/mL of penicil- 
lin, 100 [xg/mL of streptomycin, and 0.25 |xg/mL of amphotericin (Gibco, Pais- 
ley, UK) for the isolation and maintenance of neonatal rat osteoclasts. To achieve 
a basal level of resorption, the medium should be acidified by adding approx 
10 mEq/L of hydrogen ions. This can be achieved by adding 85 |iL of concen- 
trated HC1 per 100 mL of medium, as described by Goldhaber and Rabadjija 
(1987) (7) (see Note 1). 

2. Phosphate-buffered saline (PBS): For storing tissues prior to use and for remov- 
ing non-adherent cells from dentine discs. 

3. HC1: Concentrated 1 1.5 M HC1 to alter the pH of the culture medium. 

4. NaOH: 6 M NaOH to alter the pH of the culture medium. 

5. Diamond saw: Buehler "Isomet" diamond saw to cut dentine slices. 

6. Dentine slices: For bone resorption assay. 

a. Prepare the dentine slices by cutting 250-uxn thick transverse wafers from a 
block of dentine (see Note 2) using a diamond saw operating at about 60% of 
maximum speed with a moderate blade weighting. 

05/Hoebertz/51-64/F1 54 I 2/26/03, 10:45 AM 

I ■ I I I 


Isolated Osteoclast Cultures 55 

b. Soak the slices for 2 h in distilled water to reduce brittleness; and cut 5-mm 
diameter discs from the wet wafers using a standard "Rexel" single-hole paper 
punch. These discs fit neatly into the wells of 96-multiwell plates. 

c. Wash the discs extensively by sonication in multiple changes of distilled water 
and store dry at room temperature. 

d. Before use, number the discs using a graphite pencil to aid identification and 
sterilize by immersing for 1 min in 100% ethanol. 

e. Allow the discs to air-dry inside a tissue culture flow cabinet (> 30 min) and 
rinse with sterile PBS. 

7. Animals: The number of animals to be used depends on the number of treatment 
groups in the experiment. It should be borne in mind that variation within treat- 
ment groups is usually quite high in osteoclast resorption assays. Therefore, at 
least five or six replicate dentine discs should ideally be allowed for each treat- 
ment group. Generally, four rat pups, aged 2-4 d,are required for six treatment 
groups, and five animals for seven or eight groups. It is not recommended that 
more than five animals are used because the pooled, dissected bones need to be 
chopped very quickly (see Subheading 3.1.). 

8. Fixative: 2% glutaraldehyde in PBS; prepare fresh before use. 

9. Tartrate resistant acid phosphatase (TRAP) staining: leukocyte acid phosphatase 
kit (Sigma Kit 387-A). 

10. Cell removal solution: 0.25 M ammonium hydroxide. 

11. Resorption pit staining solution: 1% (w/v) Toluidine blue in 1% (w/v) sodium 
borate solution. 

12. Microscopes: A transmitted light microscope is used to count TRAP-positive 
osteoclasts and total number of cells. Number and/or area of resorption pits are deter- 
mined using brightfield reflected light microscopy (5,8). We use a Nikon "Labophot" 
2A microscope, with 100 W epi-illumination and metallurgical objectives. 

3. Methods 

3. 1. Isolation of Osteoclast from Neonatal Rat Long Bones 

1. Prior to obtaining the cells, place sterile dentine discs into the wells of a 96- well 
plate, numbered side facing down, and add 50 [iL of culture medium to each 
well. Incubate for 30 min at 37 °C. 

2. Prepare 5 mL of culture medium containing test and control substances for each 
group and add to individual wells of a six-well plate. Place in an incubator at 5% 
C0 2 -95% air at 37°C for at least 30 min. 

3. Euthanize neonatal (2—4 d) rat pups by cervical dislocation or decapitation. Cut the 
arms and legs off and dissect the long bones dissected free of muscle, connective 
tissue, and cartilage. 

4. Transfer the bones to a 6-cm diameter sterile plate (non-tissue-culture treated) 
containing 3 mL of medium. Chop the bones finely with a scalpel blade, using 
fine forceps to hold the bones steady. 

5. Create a suspension by aspirating the minced bones 10-20 times through a wide- 
mouth polyethylene 3-mL transfer pipet with the tip cut back such that the open- 
ing is about 5 mm in diameter. 

05/Hoebertz/51-64/F1 55 1 2/26/03, 10:45 AM 

I ■ I I I 


56 Hoebertz and Arnett 

6. Transfer the suspension (including the remaining small bone pieces) to a sterile 
bijou container and vortex-mix for 30 sec. 

7. Allow the mixture to settle for a few sec and — avoiding the bone fragments — 
transfer the supernatant to a fresh bijoux tube, using a 1-mL polyethylene pipet. 

8. Wash the dish and remaining bone fragments with 2 mL of culture medium, and 
vortex-mix briefly. Aspirate the supernatant and combine with the cell suspen- 
sion from step 7. 

9. Quickly add 1 00 u,L of cell suspension to each well of the 96-well plate and allow 
to settle for 45 min at 37°C (see Note 3). 

10. Carefully remove the discs (containing adherent cells) from the 96-well plates 
using fine forceps or a 19-gauge needle and rinse by dipping with gentle agitation 
in two changes of sterile PBS. 

11. Transfer to preequilibrated culture medium containing test substances or vehicle 
in a six-well plate, such that each test or control well contains 5 mL of MEM and 
five or six replicate dentine discs. 

12. Incubate for 24-28 h in a humidified atmosphere of 5% C0 2 -95% air. 

13. At the end of the experiment, measure the medium pH and pC0 2 using a clinical 
blood gas analyser, with careful precautions to prevent C0 2 loss (see Note 4). 

3.2. Fixation and Staining 

1 . On termination of the experiment, wash the dentine discs twice in PBS. 

2. Transfer to fixative for 5 min. 

3. Wash twice with PBS and leave to air-dry. 

4. Stain the wafers for TRAP-positive cells and stromal cells by following the 
directions in the kit and count TRAP-postitive multinucleated osteoclasts (see 
Note 5). 

3.3. Quantitation of Resorption 

1. Remove cells from the discs by sonication for 5 min in 0.25 M ammonium hy- 
droxide, (see Note 6). 

2. Stain the resorption pits by immersing in resorption pit staining solution for 2 min 
and allowing to air-dry (see Note 7). 

3. Count the pits by scanning the entire surface of each disc using reflected light 
microscopy and a x 20 objective. 

4. Express the results as number of pits/osteoclast or area resorbed/osteoclast and 
as pits/dentine disc (or area resorbed/dentine disc). It is usually preferable to 
normalize resorption to osteoclast numbers, as the latter may vary quite mark- 
edly within and between treatment groups (see Note 8). 

3.4. Statistics 

Depending on the data we routinely use one-way analysis of variance 
(ANOVA) or nonparametric tests (Mann-Whitney) to analyze experiments. 
Although often neglected, adjustments for multiple comparisons between treat- 
ment groups (e.g., the Bonferroni correction) are frequently needed (see Note 9). 

05/Hoebertz/51-64/F1 56 1 2/26/03, 10:45 AM 

I ■ I I I 


Isolated Osteoclast Cultures 57 

4. Notes 

1. Importance of pH: Extracellular pH is a critical factor in all bone resorption 
experiments. Rat osteoclasts are maximally activated to form resorption pits at 
pH -6.9 and resorption is essentially "switched off above pH -7.2 (Fig. 1) (4,8). 
Similar responses to extracellular acidification have been observed in all bone 
resorption systems examined to date, using cells or tissues derived from murine, 
avian or human sources (Fig. IB) (9-12). The action of many bone resorbing 
agents including parathyroid hormone (4), 1,25-dihydroxy vitamin D 3 [1,25- 
(OH) 2 D 3 ] (3), extracellular nucleotides such as ATP (14) and ADP (3), receptor 
activator of nuclear factor NF-kB ligand (RANKL) (15) is enhanced by acidifi- 
cation. Examples of the effects of pH are shown for RANKL and ADP in Fig. 
2A,B. Conversely, alkalinization attenuates the osteolytic action of parathyroid 
hormone, l,25-(OH) 2 -D 3 and prostaglandin E 2 (12). These results indicate that a 
low pH is an essential requirement for the activation process; and once this acti- 
vation has occurred, further stimulation by a wide range of bone resorbing agents 
can take place. Some tissue culture media (including MEM) are buffered to pH 
-7.20 when fully equilibrated with 5% C0 2 ; this value corresponds to normal 
interstitial pH and is considerably more acidic than blood pH (7.35-7.40). Other 
media, such as DMEM, contain higher levels of bicarbonate and consequently 
are buffered to a higher pH (-7.5 for DMEM) when equilibriated with 5% 
C0 2 .The metabolic activity of cells cultured in the medium will act to lower the 
pH further; when cell numbers are high relative to the volume of medium, this 
effect can be sufficient to acidify the medium quite rapidly, with resultant acti- 
vation of resorption pit formation. To activate resorption in a more controlled 
manner, relatively large volumes (a0.5 mL/dentine disc/24 h) of preacidified 
culture medium should be used. MEM should be acidified by the direct addi- 
tion of small amounts of concentrated HC1, which has the advantage of being 
self-sterilizing (7,8). This has the effect of reducing HC0 3 ~ concentration (i.e., 
"metabolic acidosis") and producing an operating pH close to 6.95 in a 5% C0 2 
environment (see Fig. 3), which is optimal for resorption pit formation (8). 
Further acidification in C0 2 -HC0 3 -buffered media does not enhance resorp- 
tion greatly and may ultimately reduce cell survival. Addition of HC1 has the 
effect of increasing medium chloride concentration slightly, but this does not 
appear to affect bone cell function. Culture medium can also be acidified to 
give an operating pH close to 7.0 by increasing the concentration of C0 2 to 
10% (equivalent to a partial pressure of 85 mmHg; "respiratory acidosis"), 
while HC0 3 remains constant (see Fig. 3). In organ culture systems at least, 
C0 2 acidosis is a less effective activator of resorption than HC0 3 acidosis 
(12,16). However, in contrast to the stimulatory effect of pH on osteoclast 
activity, the formation of osteoclasts from hematopoietic precursors is optimal 
at - pH 7.35-7.4 (15,17), and is inhibited at low pH. For osteoclast formation 
experiments using mouse marrow cocultures, we obtained best results when 
culture medium (MEM) was alkalinized to pH -7.35 by the addition of 10 meq/L 
of NaOH for the first 10 d (to promote osteoclast recruitment), and then replaced 

05/Hoebertz/51-64/F1 57 I 2/26/03, 10:45 AM 






V> 1-5 

• ^\ T 









o 1.0 

- \ 






' \ 

6.8 6.9 7.0 7.1 7.2 7.3 






% 1.0 



i 0.6 



7.0 7.1 7.2 7.3 7.4 7.5 


6.8 6.9 7.0 7.1 7.2 7.3 7.4 



Fig. 1. Acid activation of rat (A), chick (B), and human osteoclastoma-derived osteoclasts (C). Culture medium pH was ad- 
justed by addition of HC1 or NaOH. Rat osteoclasts are essentially "switched off" above pH 7.2, whereas chick osteoclasts retain 
some resorptive activity even at pH 7.4; maximal acid-stimulation for osteoclasts occurs at pH -6.9. Rat (and chick) osteoclasts 
were cultured as described in Subheading 3.; human osteoclasts were isolated from giant cell tumors by enzymatic and mechanical 
release and cultured under the same conditions. Values are means ± SEM (/; = 5). 












I ■ I I I 

Isolated Osteoclast Cultures 




7.43 7.41 7.40 7.41 I 6.99 6.97 6-98 6.9 

5 10 

■ physiological (blood) pH ## 

E^-3 acidified 





1 10 100 1 10 100 

RANKL (ng/ml) 

B P H 

7.079 6.825 I 7.11 4 6.824 

1 1 ' 1 1 


Control ADP (1 nM) 

Fig. 2. Synergistic stimulatory effects of low pH and RANKL (A), and low pH and 
ADP (B) on osteoclastic resorption. (A) At physiological (blood) pH of -7.4, basal 
resorption was very low and RANKL treatment caused small increases only. Com- 
bined treatment with RANKL (10 and 100 ng/mL) and low pH resulted in dramatic 
increases in resorption. (B) The stimulatory effect of ADP on resorption pit formation 
was observable clearly only when culture medium was acidified to a running pH of 
~ 6.9. Addition of 1 \iM ADP resulted in a three-fold increase in number of pits/osteo- 
clast compared to acidified control, and a 13-fold increase compared to nonacidified 
control, indicating synergism between the two stimuli (3). Significantly different from 
control in the same pH group: # #, p < 0.01, ###/?, < 0.001. Significantly different 
from the same RANKL concentration at pH 7.4: ** p, < 0.01, ***,/> < 0.001. Values 
are means ± SEM (n = 5). 

for the last 4 d with MEM acidified to pH -6.95 by addition of HC1 to ensure 
resorptive activity (Fig. 4) (10,18). 
2. Sources of dentine: Dentine can sometimes be obtained in the form of confiscated 
elephant ivory or sperm whale teeth from customs or fisheries and wildlife agen- 
cies (e.g., in the United Kingdom or United States). Dentine is a convenient osteo- 
clast substrate because it is uniform, easy to cut, and lacks features such as 
Haversian systems and osteocytes which make quantification of resorption diffi- 
cult. If bone slices are being used, they should be prepared from defatted and 
washed cortical bone from bovine femora; the slices should be transversely cut to 
reduce the likelihood of confusing in vitro resorption with endogenous features. 
We find that cortical bone is usually too brittle to permit the fabrication of uniform 
discs using a hole punch. A number of thin layer synthetic substrates are also avail- 
able. Some commercial preparations consist of hydroxy apatite sintered at high tem- 
perature onto silica discs; these have the advantage of being translucent (e.g., for 
electrophysiology or cell fluorescence work), but in our experience may impair 
osteoclast survival. Alternatively, mineralized collagen films can be prepared more 



2/26/03, 10:45 AM 

I ■ I I I 


Hoebertz and Arnett 

♦ MEM /Earle's salts /10%FCS 

■ MEM/ Earle's salts/ 10% FCS + 10 rrEq/l H* (as HCI) 

A 199 /Hank's salts/ 10% FCS 









Fig. 3. Relationship between pH, pC0 2 , and HC0 3 in tissue culture media. 

cheaply using the method of Lees et al. (19). The synthetic mineralized films also 
suffer from the disadvantage of disintegration and fragility in media acidified to 
pH 7.0 or below. 

3. Technique for plating out cell suspension: A motorized 2.5-mL multidispensing 
pipet (Rainin) is ideal for plating out the cell suspension. Delays at this stage can 
cause problems, because the cells sediment rapidly from the suspension unless it 
is continuously agitated. Care should also be taken to ensure that dentine discs 
remain seated in the base of the wells and do not float up. To compensate for 
plating errors, the suspension should be dispensed sequentially across treatment 
groups, rather than dispensing to each treatment group in turn. 

4. Measuring the medium pH: Accurate measurement of the operating pH of mam- 
malian osteoclast cultures is necessary for meaningful comparison of results from 
different laboratories. For HC0 3 -C0 2 buffered media, accurate pH measure- 
ments can be achieved only by the use of a properly standardized blood gas ana- 
lyzer. We use a reconditioned Radiometer ABL 330 blood gas analyzer, which 
was obtained at low cost (Henderson Biomedical, Beckenham, Kent, UK). The 
blood gas analyzer uses a three-electrode system to measure pH, pC0 2 and p0 2 
in a 200-u.L injected sample (cycle time ~4 min). The first medium measurement, 



2/26/03, 10:45 AM 

I ■ I I I 

Isolated Osteoclast Cultures 





7.150 7.107 7.000 6.893 

-5 5 10 

Added protons (mEq/l) 

10d alk 14d alk 

+ 4d acid 

14d acid 

Fig. 4. Effects of pH variations on osteoclast formation and resorption in mouse 
marrow cultures. (A) Mouse marrow was cultured for 10 d; "-5 meq/L added protons" 
signifies the addition of 5 meq/L OH as NaOH. Increasing the final pH from 7. 107 to 
7.150 by addition of NaOH resulted in a two-fold increase in the area covered by 
TRAP-positive osteoclasts, but a seven-fold decrease in resorption area compared to 
control, whereas addition of 5 and 10 meq/L H + resulted in a reduction of osteoclast 
formation. (B) Incubation of mouse marrow cultures for 14 d at pH 7.41 resulted in 
abundant TRAP-positive multinucleated cell formation, but almost no resorption. In 
cultures maintained in alkaline medium for 10 d, followed by 4 d in acidified medium, 
formation of TRAP-positive osteoclasts was similar to 14 d at pH 7.41, but resorption 
pit formation was increased 93-fold. Continuous incubation in acidified media (pH 
7.01) for 14 d reduced TRAP-positive multinucleate cell formation, but further 
increased the ratio of pit area/TRAP-positive area. (A) Significantly different from 

control: *, p < 0.05, 

p < 0.01. (B) Significantly different from 14-d alkaline 

medium: **,p < 0.01. (Reproduced by courtesy of Dr. M. Morrison.) 

taken immediately after removing the culture plates from the incubator, is as- 
sumed to provide a pC0 2 value that is the same for all wells and that reflects the 
actual pC0 2 during the 24-h incubation. (It is worth noting that opening the door 
of the incubator during experiments, especially prior to termination, may cause 
perturbations in C0 2 levels that affect measured pH and pC0 2 values, and possi- 
bly osteoclast function). Measured pC0 2 typically drops for each subsequent 
reading from wells in a multiwell plate, causing pH values to rise accordingly. 
The pH readings for each well are then back-corrected to the pH value associated 
with the initially measured pC0 2 value, using calibration curves previously mea- 
sured for culture medium with different bicarbonate concentrations (Fig. 3). 
5. Counting cells: Numbers of TRAP-positive multinucleated osteoclasts (three or 
more nuclei) are assessed using transmitted light microscopy and a x 20 objec- 


2/26/03, 10:45 AM 

I ■ I I I 


Hoebertz and Arnett 


Fig. 5. Typical resorption pit complex with two osteoclasts (arrowheads), viewed 
by reflected light microscopy. Scale bar = 50 u.m. 

tive. Counting should ideally be performed "blind" to the treatment group. After 
counting osteoclasts, the numbers of mononuclear cells can be assessed using 
toluidine blue staining to ensure that cells were settled at similar densities onto 
the discs and that agents tested did not have a general cytotoxic effect. Cell count- 
ing should be performed in a "blinded" manner. Reducing the time allowed for 
bone cells to sediment onto the substrate reduces the number of cells that adhere. 
When cell density is low, the effects of substance being tested are more likely to 
be direct on osteoclasts, rather than mediated via other cell types. However, low 
cell numbers can result in high levels of variation between replicates and prob- 
lematic statistics. Osteoclast counting on uniform substrates may be possible us- 
ing semi-automated methods. Although this approach is not generally feasible 
using TRAP-stained preparations (because osteoclast cytoplasm is not uniformly 
demarcated), osteoclast preparations stained using the murine monoclonal anti- 
body 23C6, directed against the vitronectin receptor, can show extremely clear 
cell definition and contrast. Our own studies have demonstrated that automated 
image analysis of low power digital images of entire 5 mm dentine discs, stained 
with antibody 23C6 to visualize human osteoclasts, is rapid and effective. 

6. Removal of cells: Sonication can remove the graphite pencil markings. In view 
of this, discs should be sonicated in a known sequence, so that identification 
numbers can be rewritten if necessary. 

7. Visualizing resorption pits: Although reflected light microscopy often yields 
adequate images of resorption pits on unstained specimens, image quality is 
improved greatly by staining because this has the effect of increasing reflectivity 



2/26/03, 10:45 AM 

I ■ I I I 


Isolated Osteoclast Cultures 63 

(Fig. 5). Depending on the microscope system used, optimal reflected light 
images are obtained using either brightfield or darkfield modes. Pits that are com- 
pletely separated by an area of unresorbed bone are counted individually, no 
matter how small; large, contiguous areas of resorption are also counted as single 
pits, as are extended "snail-trail" troughs. Although it may appear counte- 
rintuitive, the correlation between the number of discrete pits and the actual plan 
surface area resorbed is usually very close within treatment groups (17). It is 
worth bearing in mind that biochemical analysis may be possible when high lev- 
els of resorption are achieved. For example, measurement of calcium release into 
culture medium using a standard colorimetric method or a sensitive electrode 
may be possible. Such methods are inherently more efficient than measuring ana- 
tomical resorption and should yield data that reflect closely the volume of matrix 

8. Assay variability: One of the most serious problems with this assay is the high 
variability between experiments. First, osteoclast number in some culture prepa- 
rations can be low, even if the procedure is followed accurately. Second, the 
basal level of resorption can vary from experiment to experiment, perhaps 
reflecting alterations in ambient concentrations of bone resorbing agents such as 
growth factors and nucleotides. Thus, culture conditions should be kept as iden- 
tical as possible (i.e., freshness and pH of the medium; serum batches; C0 2 con- 
centration in the incubator; origin, washing, and sterilizing of the discs). If 
resorption fails to occur in cultures where authentic osteoclasts are clearly 
present, the problem can usually be rectified by acidification of the medium. To 
assess resorption area, we normally use a simple dot counting morphometry sys- 
tem: output from the reflected light microscope via a standard black and white or 
color video camera is displayed on a monitor, superimposed on which is an 
acetate sheet bearing a grid of dots. The dot grid is easily created by using a graph 
paper template that has been photocopied to the required magnification. Clearly, 
this process is more time consuming than simple pit counting and frequently it 
contributes little extra information. It is possible to measure the volume resorp- 
tion pits using SEM or confocal microscopy (6), but this requires specialized and 
expensive equipment. Pit volume can also be estimated by measuring pit depth 
and area, using reflected light microscopy (the fine focus control is usually cali- 
brated in microns) and assuming that pits approximate to hemispheres; this 
method is not suited to determining the volume of individual pits to very high 
accuracy, but provides useful comparative data when multiple pits are measured. 

9. Because of interassay variation, statistical comparisons should be performed only 
within one assay, and not between different assays. Variability within assays is 
likely to be high when cell numbers are low (e.g., < 10 osteoclasts/disc); the 
inherent "noise" in osteoclast resorption assays means that they are best suited 
for studying large, robust effects. 


1. Boyde, A., Ali, N. N., and Jones, S. J. (1984) Resorption of dentine by isolated 
osteoclasts in vitro. Br. Dent. J. 156, 216-220. 

05/Hoebertz/51-64/F1 63 1 2/26/03, 10:45 AM 

I ■ I I I 


64 Hoebertz and Arnett 

2. Chambers, T. J., Revell, P. A., Fuller, K., and Athanasou, N. (1984) Resorption of 
bone by isolated rabbit osteoclasts. /. Cell Sci. 66, 383-399. 

3. Hoebertz, A., Meghji, S., Burnstock, G., and Arnett, T. R. (2001) Extracellular 
ADP is a powerful stimulator of bone resorption: evidence for its signaling through 
the P2Y! receptor. FASEB J. 15, 1 139-1 148. 

4. Arnett, T. R. and Dempster, D. W. (1986) Effect of pH on bone resorption by rat 
osteoclasts in vitro. Endocrinology 119, 119-124. 

5. Walsh, C. A., Beresford, J. N., Birch, M. A., Boothroyd, B., and Gallagher, J. A. 
(1991) Application of reflected light microscopy to identify and quantitate 
resorption by isolated osteoclasts. /. Bone Miner. Res. 6, 661-671. 

6. Boyde, A. and Jones, S. J. (1991) Pitfalls in pit measurement. Calcif. Tissue Int. 
49, 65-70. 

7. Goldhaber, P. and Rabadjija, L. (1987) H + stimulation of cell-mediated bone 
resorption in tissue culture. Am. J. Physiol. 253, E90-E98. 

8. Arnett, T. R. and Spowage, M. Modulation of resorptive activity of rat osteoclasts by 
small changes in extracellular pH near the physiological range. Bone 18, 277-279. 

9. Morrison, M. S. and Arnett, T. R. (1997) Effect of extracellular pH on resorption 
pit formation by chick osteoclasts. /. Bone Miner. Res. 12, S290 (Abstr.). 

10. Morrison, M. S. and Arnett, T. R. (1998) pH effects on osteoclast formation and 
activation. Bone 22, 30S (Abstr.). 

11. Hoebertz, A., Nesbitt, S. A., Horton, M. A., and Arnett, T. R. (1999) Acid activa- 
tion of osteoclasts derived from human osteoclastoma. /. Bone Min. Res. 14, 
SI 049 (Abstr.). 

12. Meghji, S., Morrison, M. S., Henderson, B., and Arnett, T. R. (2001) pH-depen- 
dence of bone resorption: mouse calvarial osteoclasts are activated by acidosis. 
Am. J. Physiol. 280, El 12-E1 19. 

13. Murrills, R. J., Dempster, D. W., and Arnett, T. R. (1998) Isolation and culture of 
osteoclasts and osteoclast resorption assays, in Methods in Bone Biology (Arnett, 
T. R. and Henderson, B., eds.), Chapman and Hall, London, pp. 64-105. 

14. Morrison, M. S., Turin, L., King, B. F., Burnstock, G., and Arnett, T.R. (1998) 
ATP is a potent stimulator of the activation and formation of rodent osteoclasts. /. 
Physiol. 511, 495-500. 

15. Zanellato, N., Hoebertz, A., and Arnett, T. R. (2000) Low pH is a key requirement 
for osteoclast stimulation by RANKL. /. Bone Miner. Res. 15, S387 (Abstr.). 

16. Bushinsky, D. A. (1987) Net calcium efflux from live bone during chronic meta- 
bolic, but not respiratory, acidosis. Am. J. Physiol. 256, F836-F842. 

17. Arnett, T. R. and Dempster, D. W. (1987) A comparative study of disaggregated 
chick and rat osteoclasts in vitro: effects of calcitonin and prostaglandins. Endo- 
crinology 120, 602-608. 

18. Shibutani, T. and Heersche, J. N. (1993) Effect of medium pH on osteoclast activ- 
ity and osteoclast formation in cultures of dispersed rabbit osteoclasts. /. Bone 
Miner. Res. 8,331-336. 

19. Lees, R. L., Sabharwal, V. K., and Heersche, J. N. (2001) Resorptive state and 
cell size influence intracellular pH regulation in rabbit osteoclasts cultured on 
collagen-hydroxy apatite films. Bone 28, 187-194. 

05/Hoebertz/51-64/F1 64 I 2/26/03, 10:45 AM 

I ■ I I I 


Primary Isolation and Culture 
of Chicken Osteoclasts 

Patricia Collin-Osdoby, Fred Anderson, and Philip Osdoby 

1 . Introduction 

Bone is a dynamic tissue that is continually remodeled throughout life. Such 
remodeling is carried out by the coordinated actions of two bone cell types: 
bone-resorbing osteoclasts (OCs) which are uniquely capable of dissolving and 
removing a small volume of bone, and bone-forming osteoblasts that subse- 
quently fill in these lacunae or pits with new bone tissue. Whereas osteoblasts 
originate from mesenchymal cell precursors, OCs derive from hematopoietic 
precursors related to monocytic cells that are present both in the bone marrow 
and peripheral circulation. In response to specific hormonal or local signals 
provided by osteoblasts, stromal cells, or other cells within the bone marrow 
microenvironment, OCs precursors fuse and differentiate into large multinucle- 
ated cells expressing characteristic morphological features, membrane polar- 
ization, adhesion molecules, ion pumps, enzyme activities, and antigenic 
profiles (1-3). Most importantly, they develop a capacity for bone pit resorp- 
tion, the unique and defining functional attribute of OCs. Bone resorption and 
formation are normally carefully balanced processes in adults. However, in 
various diseases or pathological conditions, an imbalance exists such that the 
number of OCs, number of resorption sites initiated, and/or rates of remodel- 
ing are altered, thereby resulting in either too much or too little bone turnover. 
Excessive bone loss occurs in many clinically relevant disorders that affect 
millions of people, including postmenopausal osteoporosis, rheumatoid arthri- 
tis, periodontal disease, tumor-associated osteolysis, and orthopedic implant 
loosening (4-7). It is therefore important to decipher the complex signals that 
control OC bone resorption to further our understanding and provide a rational 
basis for the design of novel therapeutic or preventative strategies to combat 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


06/Collin/65-88/F1 65 1 2/26/03, 10:45 AM 

I ■ I I I 


66 Collin-Osdoby et al. 

bone loss. In this regard, isolated primary cultures of in vivo formed OCs have 
proven to be an invaluable tool for investigating the characteristics, function, 
and regulation of OCs. In particular, many insights have been achieved through 
the use of avian OCs, cells that are highly active in their resorption of bone and 
are readily isolated in abundance from young chickens fed a low-calcium diet 
(8). Following their enzymatic release from the bones harvested from such 
animals, OCs can be partially purified by density gradient (Percoll) sedimenta- 
tion owing to their large size, and further enriched by rapid capture with mag- 
netic beads precoupled with an antibody that specifically recognizes OCs (9). 
OCs isolated via either procedure can be cultured and analyzed for biochemi- 
cal, immunological, physiological, and functional properties, as well as modu- 
lator responses. Procedures for some of the most commonly used assays are 

2. Materials 

2.1. Tissue Culture Medium, Solutions, and Supplies 

All media and solutions should be prepared with glass distilled water. 

1. Culture medium: Sterile a-mimimum essential medium (a-MEM) supplemented 
with 5% fetal bovine serum (FBS, Gibco) and 2.5% antibiotic/antimycotic (a/a, 
Gibco); store at 4°C and prewarm to 37°C for use with cells. 

2. Hanks' balanced salt solution (HBSS, Gibco): Dissolve one packet in 990 mL of 
water, add 10 mL of a/a and 3.5 g of NaHC0 3 , check that the pH is 7.2, sterile- 
filter one batch, and prepare another to store at 4°C without sterilization. 

3. Moscona's low bicarbonate (MLB): Add 8 g of NaCl, 0.2 g of KC1, 50 mg of 
NaH 2 P0 4 , 0.2 g of NaHCO,, 2 g of dextrose, 10 mL of a/a, and 990 mL of water; 
check that the pH is 7.2; and sterile-filter. 

4. Moscona's low bicarbonate-EDTA (MLBE): Dissolve 1 g of EDTA in 15 mL of 
1% KOH, add to 1 L of MLB, check that pH is 7.2, and sterile-filter. 

5. Phosphate-buffered saline (PBS): Add 9 g of NaCl, 0.385 g of KH 2 P0 4 , and 1.25 g 
of KHP0 4 per liter of water (final volume); adjust pH to 7.2 using 10 N NaOH. 

6. Collagenase: Prepare 0.5 mg/mL of stock solution in HBSS, store in aliquots at 
-20°C; dilute two parts of thawed stock solution with one part of MLB for use. 

7. Trypsin: Prepare 1% stock (1 g in 100 mL) solution in MLB, store in aliquots at-20°C; 
dilute 1 1 .25 mL of stock with 37.5 mL of MLBE and 201 .5 mL of MLB for use. 

8. Percoll: Dissolve one packet of HBSS powder (for 1 L) in 640 mL of water, add 
0.35 g of NaHC0 3 and 10 mL of a/a, and sterile-filter. For 35% Percoll, mix 65 mL 
of this solution with 35 mL of Percoll (Pharmacia). For 6% Percoll, mix 83 mL of 
the HBSS solution with 17 mL of 35% Percoll-HBSS. Adjust the pH of the 35% 
and 6% Percoll solutions to 7.2 and sterile-filter. 

9. Heparin: Sterile solution (1000 U/mL, Pharmacia, store at 4 C C). 

10. Trypan blue: Dissolve 0.4 g of trypan blue dye in 100 mL of water and sterile 

06/Collin/65-88/F1 66 1 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 67 

11. 1% Paraformaldehyde in HBSS (PF-HBSS): Preheat 100 mL of HBSS on 
a hot plate to 60°C in a Pyrex beaker (monitor with a thermometer), move 
the beaker to a stir plate, add 1 g of PF, cover with foil to contain the 
vapors (keep the thermometer in place and briefly move the beaker back 
to the hot plate if the temperature falls below 50°C), slowly stir with a 
magnetic bar, and add three or four drops of 10 N NaOH just to dissolve. 
Let cool, filter through Whatman no.l paper into a brown glass bottle, and 
store at 4°C. 

12. Protease inhibitor cocktail for cell pellet storage: Prepare an inhibitor stock solu- 
tion A by dissolving 10 mg each of leupeptin, chymostatin, antipain, and pepstatin 
A in 1 mL of dimethyl sulfoxide, add 400 trypsin inhibitory units of aprotinin, 
and store this lOOOx cocktail in 0.1-mL aliquots at -20°C. Mix 10 \xL of this 
inhibitor stock solution A with 10 uL of a 1% stock solution of phenyl- 
methylsulfonyl fluoride (PMSF) in ethanol (store at room temperature), 1.25 mg 
of W-ethylmaleimide (NEM), 1.56 mg of benzamidine, and 10 mL of HBSS to 
yield an inhibitor stock solution B. Store this at -80°C and overlay one drop (-50 uL) 
on top of each cell pellet to be stored frozen. 

13. 350- and 1 10-um Nitex filters: Sheets of Nitex (Tetko, Kansas City, MO) mesh 
are cut into squares larger (-50%) than the opening of a stackable plastic beaker, 
a filter square is stretched over the beaker, and the filter is secured in place by 
fitting in a ring (-1 inch deep) made from a second plastic stackable beaker whose 
lower three-quarter portion has been cut off (filter squares can be washed well 
and reused). 

2.2. Preparation of Antibody-Conjugated Magnetic Beads 

1. Magnetic polystyrene beads: 0.45 urn diameter, covalently conjugated with 
affinity purified sheep anti-mouse IgG (Dynal Inc., store at 4°C). 

2. Mouse monoclonal antibody (MAb) to OC-specific antigen: See Note 1. 

3. Rotary mixer: To fit microcentrifuge tubes. 

4. Magnet: Dynal Inc. or other 20-lb pull magnet. 

2.3. Fixation and TRAP Staining 

Although not fully specific for OCs, high tartrate-resistant acid phosphatase 
(TRAP) activity is a characteristic of OCs that is upregulated in OC develop- 
ment and important for their resorption of bone (10). 

1. 1% paraformaldehyde in HBSS: See Subheading 2.1., step 11. 

2. For TRAP staining: Prepare the following stock solutions and mix just before use 
(or purchase a staining kit (cat. no. 386) from Sigma and follow the 
manufacturer's instructions; see Note 2). 

Solution A: Naphthol AS-BI phosphoric acid (12.5 mg/mL) in dimethyl 

formamide; store at -20°C. 

Solution B: 2.5 M acetate buffer, pH 5.2; store at 4°C. 

Solution C: 0.67 M tartrate solution, pH 5.2; store at 4°C. 

06/Collin/65-88/F1 67 I 2/26/03, 10:45 AM 

I ■ I I I 


68 Collin-Osdoby et al. 

a. Mix 0.4 mL of solution A, 0.4 mL of solution B, 0.4 mL of solution C, and 8.8 mL 
of deionized water (preheated to 37°C) in a 50-mL polypropylene tube and 
vortex-mix well. Wrap the tube in foil. 
b.To this add 3 mg of Fast Garnet GBC salt, quickly vortex to mix well, and 
filter the solution through Whatman no. 1 paper into a new foil- wrapped 50-mL 
polypropylene tube. Use immediately. 
3. For a general stain: Use Difquik (eosin Y, azure A, and methylene blue, Criterion 
Sciences) as recommended by the manufacturer. 

2.4. Fixation and Immunostaining 

For immunostaining, prepare the following solutions: 

1. Blocking solution: 1% Bovine serum albumin (BS A) and 10% horse serum in PBS. 

2. Monoclonal (MAb) or polyclonal (PAb) antibodies: Directed against OC anti- 
gens and appropriately diluted (typically 1:100 to 1:500 of 1 mg/mL stocks) in 
blocking solution just prior to use. 

3. Biotinylated secondary antibodies: Directed against the primary antibody and 
appropriately diluted (typically 1:200 to 1:500) in blocking solution just prior to 
use (see Note 3). 

4. Glycerol-buffered mounting medium: For example, EM Sciences, 80% glycerol 
in PBS, Vectashield (Vector), or Citifluor; store at 4°C. 

For fluorescence immunostaining: 

5. Streptavidin conjugated with a fluorescent label (fluorescein isothiocyanate 
[FITC], Texas red, or similar): Appropriately diluted (typically 1:1000 or more) 
in PBS (without serum) just prior to use. 

6. 4',6-Diamidino-2-phenylindole (DAPI): (Molecular Probes), prepare 100 |xg/mL 
of stock solution in water, store at 4°C, and dilute stock 1 :300 in HBSS for use in 
fluorescent nuclear staining. 

For colorimetric immunostaining: 

7. Streptavidin conjugated with |3-galactosidase: Appropriately diluted (typically 
1:100) in buffer A (see step 8). 

8. Buffer A: 0.1 M Sodium phosphate, pH 7.2, containing 1.5 mM magnesium chlo- 
ride, 2 mM (3-mercaptoethanol, and 0.05% sodium azide; store at 4°C and warm 
to room temperature before use. 

9. Buffer B: 10 mM Sodium phosphate, pH 7.2, containing 150 mM sodium 
chloride, 3 mM potassium ferricyanide, 3 mM potassium ferrocyanide, and 
1 mM magnesium chloride; store at 4°C and warm to room temperature 
before use. 

10. Substrate solution (0.42 mg/mL of X-gal in buffer B): Prepare a stock solution of 
X-gal (21 mg/mL) in dimethyl formamide to store at -20°C (e.g., in a parafilm 
sealed, foil wrapped glass tube), and dilute this stock solution 1:50 in buffer B to 
prepare the required fresh substrate solution. 

06/Collin/65-88/F1 68 1 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 69 

2.5. Preparation of Devitalized Bone or Ivory Discs for Bone Pit Re- 
sorption Studies 

1 . Ivory is obtained either through donation from a local zoo or, in the United States, 
the Federal Department of Fish and Wildlife Services (or similar department in 
other countries; see also the chapter by Nesbitt and Horton, this volume). Bovine 
cortical bone is obtained from a local slaughterhouse. Segments of ivory and 
bovine cortical bone are thoroughly cleaned and washed (multiple HBSS and 
70% ethanol rinses), sliced into small chunks, and then reduced to rectangular 
0.4-mm thick sheets using a low-speed Isomet saw (Buehler, Lake Bluff, IL). 

2. The sheets are rinsed three times with 70% ethanol, incubated in 70% ethanol 
overnight, and then washed for several hours in HBSS before circular discs are 
cut using a 5-mm paper punch. 

3. The discs are soaked repeatedly in 70% ethanol in sterile 50-mL tubes (alcohol 
changes can be gently poured off because the discs tend to stick to the side of the 
tube), and stored in 70% ethanol at -20°C. 

4. For experimental use, the required number of discs are removed from the tube 
using alcohol-presoaked tweezers (to maintain sterility) in a tissue culture hood, 
transferred to a fresh sterile 50-mL polypropylene tube, rinsed extensively by 
inversion and mild shaking at least three times with -40 mL of sterile HBSS per 
wash, and the discs transferred using sterile tweezers into culture wells or dishes 
containing sterile HBSS for 3-24 h of preincubation in a tissue culture incubator 
prior to the plating of cells. HBSS is removed only immediately before the discs 
are to be used so that they do not dry prior to OC seeding. 

2.6. Preparation of Gold-Coated Glass Coverslips for Phagokinetic 
Motility Studies (see Note 4) 

This procedure is a modified version of the gold coverslip motility assay 
reported by Owens and Chambers (11). Glass coverslips are precoated with a 
thin layer of gelatin to enhance attachment and homogeneous coverage of the 
gold coating. All steps are performed in a sterile hood, using sterile reagents 
and supplies, and more coverslips (10-50%, depending on the skill you develop 
for this procedure) should be coated than you expect to need in the experiment. 

1. Place a 2% gelatin solution prepared in deionized water (which can be stored at 
room temperature) into a 37°C water bath. 

2. Warm up two 24-well tissue culture dishes to 37°C (e.g., in an incubator). 

3. Place glass coverslips, sterile tweezers, and one or two 100-mm Petri dishes con- 
taining a piece of Whatman filter paper into the hood. 

4. Fill a single well of one prewarmed 24-well dish with warm 2% gelatin solution 
and dip each coverslip individually into the well using sterile tweezers. 

5. Briefly drain each coverslip by touching it against the side of the well, and place 
it gelatin side up onto the filter paper in the open 100-mm dish to dry in the hood 
for at least 2 h. 

06/Collin/65-88/F1 69 1 2/26/03, 10:45 AM 

I ■ I I I 


70 Collin-Osdoby et al. 

6. Using tweezers, move each coverslip into one well of a sterile 24-well dish in the 

7. Prepare the gold coating solution (19.0 mL) in the hood by adding 11.0 mL of 
sterile deionized water, 4.44 mL of 0.2% gold chloride (in sterile water), and 
3.36 mL of sterile 65.2 vM sodium carbonate into a 100-mL Pyrex beaker (foil 
covered and presterilized). Heat the solution on a hot plate (this can be performed 
outside the hood after the foil cover is replaced) just to boiling. 

8. Place the hot beaker back into the hood and add 1.8 mL of a sterile solution of 
0.1% paraformaldehyde in water. 

9. Allow the gold solution to cool to 60°C in the hood. Monitor the temperature 
(this is critically important to achieve good gold coating) with an alcohol 
prewiped thermometer. 

10. When the solution has cooled to 60°C, pipet 1 mL of the gold solution on top of 
each gelatin precoated coverslip and then place the 24-well dish into the refrig- 
erator for 1 h. 

11. Remove the excess solution, gently rinse the coverslips with HBSS twice, and 
place each coverslip onto filter paper in a 100-mm dish in the sterile hood to dry 
(several hours to overnight). 

12. Repeat steps 7-11 to ensure adequate and even coverage of the gold particles on 
the coverslips. 

13. After the second gold coating, the coverslips are stored on the filter paper in a 
100-mm dish in the hood for up to a few days prior to their experimental use. 
Check one or two coverslips the next day by placing into a 24-well dish with 
HBSS for at least 1 h to verify that the gold coating does not lift up and that it is 
sufficiently dark and evenly coated when viewed under the microscope (other- 
wise tracks produced by OCs will be hard to evaluate). If the coating lifts up, 
check a few other coverslips from that batch and discard them all if they fail this 
test. If the gold coating is too sparse, repeat steps 7-11 for a third time. 

14. Before use, preincubate all of the coverslips for at least 1 h in HBSS early on the 
day of cell plating and plan to use only those that exhibit a firm, even, and dark 
gold coating. 

3. Methods 

3.1. Isolation of Osteoclasts from Calcium-Deficient Chicks 
(see Note 5) 

White Leghorn chick hatchlings are initially fed a normal diet for 4-6 d and 
then placed on a low calcium diet (0.15-0.25% calcium, analyzed before the 
feed is shipped from Purina) for at least 28 d (8). Typically, 15 chicks are used 
for each OC preparation and three people assist in the dissection and initial 
steps of the protocol to minimize the time involved until the cells are isolated 
and plated (even an additional hour can affect the ultimate OC yield and 
viability; see Note 6). Animals are handled and euthanized in accordance with 
approved rules and procedures for the Institutional Animal Care and Use Com- 

06/Collin/65-88/F1 70 1 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 71 

mittee and standards approved by the National Institutes of Health Guidelines 
for the Care and Use of Experimental Animals (or similar appropriate author- 
ity for countries other than the United States). 

1. Just prior to dissection, forceps, tweezers, and scissors should be placed into 
beakers of 70% alcohol and all buffers prechilled on ice. 

2. Readjust the pH of MLB if necessary. 

3. Fill several ice containers and place two 100-mm Petri dishes with HBSS on ice. 

4. Several people wearing alcohol-rinsed gloves should each remove a group of 
birds immediately after they are euthanized (do not delay), alcohol squirt the 
wings and legs of each bird just prior to its dissection, rapidly remove the tibiae 
and humeri using the alcohol-soaked scissors and forceps, clean off extraneous 
soft tissue without removing the bone ends (which are replete with OCs), and 
place the bones into one of the two HBSS-filled Petri dishes on ice. 

5. When a number of bones have accumulated in the first HBSS dish on ice, this 
dish is given to one person to remove the marrow from the bones, while the other 
people continue to dissect bones from the remaining birds and place them into 
the second HBSS dish on ice. 

6. Remove the marrow from each bone by gripping the bone with alcohol-soaked 
tweezers over another 100-mm dish containing MLB, poking several small holes 
in each end of the bone using a 3-mL syringe fitted with an 18-guage needle, and 
quickly flushing the marrow out by repeatedly inserting the tip of the syringe 
filled with MLB (from the dish) into the end of the bone and pushing this fluid 
through the marrow cavity into the lower dish (if the marrow is to be cultured, 
this step is done in a sterile tissue culture hood by one individual working on 
some of the bones while the others are dissecting more bones out of the birds). 

7. Carefully flip the bone over and repeat step 6 by flushing MLB several more 
times through holes poked into the other end of the bone. Place this bone back 
into the original (nonmarrow) dish of HBSS. Repeat for the next bone, each time 
flushing MLB through both ends of the bone before proceeding to the next bone. 

8. After all of the marrow has been extruded from the bones, remove any remaining 
bits of extra tissue carefully from the bones and place the bones into eight 50-mL 
tubes each filled with 40 mL of HBSS. Shake gently by hand to wash the bones 
(-30 sec). 

9. Divide and place the bones into two new dishes of HBSS on ice to split each bone 
lengthwise using sterile scissors while keeping the bones submersed in HBSS. 

10. Transfer the split bones into eight 50-mL polypropylene tubes containing 40 mL 
of HBSS each, shake vigorously for 30 sec, and pass the supernatants sequen- 
tially through 350- and 1 lO-jim Nitex filters fitted over plastic beakers set on ice 
(be sure the filters are tightly fitted so that the cell suspensions do not leak into 
the bottom beaker unfiltered). 

11. Refill each of the eight tubes containing bones with 40 mL of MLB, shake, and 
filter these supernatants through the same 350- and 1 10-u.m filters into the bea- 
kers containing the first filtrates on ice. 

06/Collin/65-88/F1 71 I 2/26/03, 10:45 AM 

I ■ I I I 


72 Collin-Osdoby et al. 

12. Dispense the final filtered solutions into 50-mL centrifuge tubes on ice and cen- 
trifuge at 2l0g for 10 min at 4°C. This pellet represents a crude fraction contain- 
ing the majority of the nonviable OCs (which tend to be much larger in size on 
average than the surviving OCs) and a minor proportion of viable OCs. Because 
it provides a valuable source of OC material for enzyme-linked immuno- 
sorbentassay (ELISA), sodium dodecyl sulfate-polyacrylamide gel electrophore- 
sis (SDS-PAGE), Western blotting, and other biochemical assays, it is routinely 
stored at -80°C as one or two cell pellets overlaid with a drop of protease inhibi- 
tor cocktail solution B. 

13. To obtain the viable OC fraction for cell culture studies, incubate the bones in 
eight 50-mL tubes with 35 mL of 0.333 mg/mL of collagenase in HBSS-MLB 
for 30 min at 37°C (e.g., stationary in a water bath). 

14. Gently shake the tubes by hand, and discard this solution. Then incubate the bones 
in eight tubes containing 35 mL of MLB each for 15 min to rinse, use tweezers to 
pick the bones out of these tubes and transfer them to eight tubes containing 35 mL 
of 0.045% trypsin in MLB-MLBE, and incubate for 30 min. at 37°C (stationary 
in a water bath) to detach viable OCs from the bone surfaces. 

15. Shake the bones vigorously for 3 min., and pass the resulting cell suspension 
through a 350-u.m Nitex filter into a plastic beaker on ice containing 1 mL of hep- 
arin (1000 U, to reduce clotting) and 5 mL of FBS (to inhibit further trypsin action). 

16. Immediately refill the tubes containing the bones with 20 mL of MLB, shake 
vigorously for another 3 min., and filter this cell suspension through the 350-[xm 
Nitex into the beaker on ice containing the first shaken suspension from step 15. 

17. Again, refill the tubes containing the bones with 20 mL of MLB, shake vigor- 
ously for 1 min., and filter this solution through the 350-|xm Nitex into the beaker 
on ice containing the first and second shaken cell suspensions. 

18. Pour half of the filtered solution from steps 15-17 through one 110-u.m Nitex 
filter into a beaker on ice and the other half through another similar filter into a 
second beaker on ice (two filters, or sometimes three, are recommended because 
they get clogged easily). Dispense the filtrates into twelve 50-mL centrifuge tubes 
held on ice. 

19. Centrifuge the filtered cell suspensions at 300g for 10 min at 4°C. Gently pour out 
the lipid pad and supernatant, and wipe out lipid and matrix material clinging to the 
side of the tube with a clean tissue before inverting each tube. Resuspend each 
pellet, using a 10-mL wide-bore pipet, in 2-5 mL of chilled MLB (see Note 7). 

20. Transfer the OC suspension to six new 50-mL tubes, add 0.1 mL of heparin per 
tube, fill the tubes to 50 mL with chilled MLB, and invert to mix. 

21. Centrifuge the cell suspensions again as in step 19 to wash the cells. 

22. Discard the supernatants and resuspend the cell pellets (see Subheading 3.1.1.). 
If OCs are to be cultured, sterile techniques and solutions should be initiated at 
this point for Percoll fractionations. 

3.1.1. Percoll Purification of Osteoclasts 
Use sterile solutions and techniques throughout. 

06/Collin/65-88/F1 72 I 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 73 

1. Resuspend each of the six OC pellets in 5 mL of ice-cold 35% Percoll. 

2. Combine the resuspended pellets into one new tube and add 0.6 mL of heparin. 

3. Vortex-mix the tube briefly at low speed and divide the cell suspension into four 
50-mL tubes. 

4. Raise the volume of each tube to 10 mL with additional 35% Percoll. 

5. Slowly overlay each tube with 3.0 mL of ice-cold HBSS (try not to deform the 

6. Centrifuge the tubes in a swinging bucket rotor at 440g for 20 min at 4°C. 

7. Carefully remove the tubes without disturbing the gradients, and slowly with- 
draw the interface and top 5-8 mL with a pipet (see Note 7). 

8. Transfer this interface/top solution into four new 50-mL tubes on ice containing 
25 mL of HBSS, and then fill the tubes with additional ice-cold HBSS to 50 mL. 
Discard the tubes with remaining pellets. 

9. Centrifuge at 300g for 10 min at 4°C, and discard the supernatant. 

10. The 35% Percoll fractionated OC obtained can either be used for immuno- 
magnetic purification (see Subheading 3.1.2.) or purified further by 6% Percoll 
fractionation as described in steps 11-22. 

11. Set up four tubes containing 10 mL of 6% Percoll on ice. 

12. Resuspend each of the four OCs pellets from the 35% Percoll separation thor- 
oughly in 3 mL ice-cold HBSS using a 10-mL pipet. 

13. Combine the suspensions and briefly vortex-mix (if clumping is a problem, add 
0.12 mL of 1000 U/mL of heparin). 

14. Slowly overlay 3-3.5 mL of this suspension on top of each of the four 6% Percoll 
gradient tubes. 

15. These tubes are left undisturbed standing upright on ice for 1 h to allow OCs to 
penetrate the Percoll layer (under Ig). 

16. Remove the top 4 mL from each tube and discard. 

17. Combine the bottom fractions pairwise and dilute with ice-cold HBSS to 50 mL each. 

18. Centrifuge the cell suspensions at 300g for 10 min at 4°C. 

19. Resuspend each of the two OC pellets in 5 mL of a-MEM medium, combine, mix 
gently, and analyze by withdrawing 0.1 mL into a microcentrifuge tube contain- 
ing 0.1 mL of 0.4% trypan blue to assess immediately OC yield, viability (un- 
stained cells), and purity using a hemocytometer. 

20. Meanwhile, centrifuge the OC suspension again at 300g for 10 min at 4°C. To 
culture, resuspend the OC pellet in prewarmed culture medium and disperse into 
sterile tissue culture dishes or wells. 

Typically, enrichments of at least 40% on a per cell basis (>80% on a per 
nucleus basis) are routinely achieved for OCs after 35% Percoll fractionation, 
and yields of 1-3 million OCs exhibiting > 85% viability for OCs are obtained 
following 6% Percoll fractionation. Further enrichment of the 6% Percoll OCs 
population can be accomplished readily by allowing these cells to attach to 
bone or ivory in culture for 2.5-3 h, after which the unbound cells are removed 
and the adherent OCs are gently washed once or twice with fresh medium 
before further culture. 

06/Collin/65-88/F1 73 1 2/26/03, 10:45 AM 

I ■ I I I 


74 Collin-Osdoby et al. 

3. 1.2. Immunomagnetic Purification of Osteoclasts (see Note 8) 

Partially purified OCs can be further enriched after the 35% or 6% Percoll 
fractionation step via immunomagnetic bead capture (9). However, signifi- 
cantly higher OC yields are obtained if immunomagnetic capture is performed 
on 35%, rather than 6%, Percoll-separated populations. The last steps of bead 
preparation (steps 1-4 below) should be timed so that the beads are ready to 
add as soon as the OCs have been separated on the 35% Percoll gradients. 

1 . Just before they are to be used for OC enrichment, magnetically sort the beads that 
have been coupled with an anti-OC MAb to remove the MAb coupling solution. 

2. Wash the beads three times by gentle resuspension in PBS (~1 mL) and magnetic 

3. Incubate the beads for 30 min. with 250 uL of sterile 1% FBS in PBS to block 
nonspecific attachment of the beads to the cells. 

4. Wash the beads gently three times with PBS, and resuspend in 200 uL of PBS in 
preparation for addition to the 35% Percoll-fractionated OCs. 

5. At this point, the four OC pellets from the 35% Percoll gradient (see Subheading 
3.1.1., step 10) are resuspended in a 50-mL polypropylene tube in a total volume 
of 6 mL of ice-cold HBSS. 

6. Add the MAb-coupled magnetic beads from step 4 (200 pL) and swirl the tube 
gently to mix cells and beads quickly (see Note 9). 

7. Place the tube into a container (bucket) of ice at a -45° angle with the bead-cell 
mixture clearly visible from the top. 

8. Place the container with the bead-cell mixture on a rotary shaker and adjust the 
speed to mix the beads and cells slowly for 30 min (see Note 9). 

9. Remove from the rotary shaker. Stand the tube upright in the ice container. Push 
a magnet down into the ice and tightly against the lower vertical portion of the 
tube. Let stand for ~5 min. to draw bead-bound OCs over to the magnet side of 
the tube. Using a pipet, slowly withdraw the unbound (nonbead) cell supernatant 
and transfer it to a new 50-mL tube on ice (this is resorted later to capture any lost 
beads bound with cells). 

10. Move the tube away from the magnet. Resuspend the bead-bound cells in 40 mL 
of ice-cold sterile HBSS to wash, invert several times to mix gently, then place 
the tube back against the magnet in ice. Incubate undisturbed for 5 min. to cap- 
ture bead-bound OCs and then remove the wash supernatant with a pipet to 
another tube (to resort later for lost beads). 

11. Repeat step 10 twice more to wash the bead-bound OCs a total of three times. 

12. To recover any bead-bound OCs remaining in the original unbound cell superna- 
tant (step 9) or lost during the wash steps (steps 10 and 11), incubate each of 
these tubes against the magnet for 5 min. on ice, remove and discard the superna- 
tants, and resuspend any additional bead-bound OCs that were captured in a small 
volume of HBSS and add back to the main sample of bead-bound OCs. 

13. Resuspend the final collection of bead-bound cells in 2-5 mL of HBSS, transfer 
0.1 mL to a microcentrifuge tube, add 0.1 mL of 0.4% trypan blue to the 

06/Collin/65-88/F1 74 I 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 75 

microcentrifuge tube, and immediately assess this sample for OC yield, viability 
(unstained cells), and purity in a hemocytometer. 
14. Sort the remaining 2-5 mL of immunocaptured OCs with a magnet and either 
resuspend the cells in medium for culture, immediately extract RNA, prepare 
protein lysates, or use the cells in some similar fashion. 

Typically, 5- to 10-fold greater OC purity is achieved with MAb 12 IF 
immunomagnetic affinity capture in comparison with 6% Percoll density gra- 
dient fractionation, and immunomagnetic OC enrichments of up to 90% on a 
per cell basis and over 98% on a per nucleus basis are achieved. 

3.2. Osteoclast Culture 

3.2.1. Percoll Purified OCs 

1. Resuspend purified OCs from the 6% Percoll separation in 5 mL of culture 
medium and immediately plate out at: 

a. 0.5 mL per well (-100,000 OCs) in 10 wells of a 24- well dish, with or without 
a glass coverslip in the bottom of the well and/or two to four sterile discs of 
bone or ivory per well (see Subheading 2.5.), or 

b.0.2 mL per well (-50,000 OCs) in 20 wells of a 48-well dish, with or without 
one sterile disc of bone or ivory (see Note 10). 

2. To enrich further for OCs on bone or ivory, change the medium after 2-3 h of 
incubation, and then add modulators in fresh medium. Otherwise, change the 
medium after 16 h of incubation, and add modulators in fresh medium. 

3. Culture for the designated period of time, typically 1-2 d (see Note 11). 

3.2.2. Immunomagnetically Purified OCs (see Note 8) 

Because the yield of OCs is lower following immunomagnetic purification 
than Percoll density gradient separation, highly purified immunomagnetic OC 
populations are considered most useful for confirming in a limited number of 
experiments that biochemical or functional effects observed using Percoll frac- 
tionated OCs can be directly attributed to OCs. 

1 . Culture immunomagnetically purified OCs on bone, ivory, glass, or plastic in the 
presence or absence of modulators for up to several d as described in Subheading 
3.2.1. above. 

2. Although the temperature is kept at or below 4°C to prevent OCs from phagocy- 
tosing the beads bound to their outer surface during immunomagnetic capture, 
these beads will be internalized within min once the cells are exposed to a higher 
temperature (see Note 12). 

3.3. Assay Techniques 

3.3.1. Morphology and Ultrastructure 

Standard protocols can be used to evaluate the morphological and ultrastruc- 
tural characteristics of isolated chick OCs. When viewed by light microscopy, 

06/Collin/65-88/F1 75 1 2/26/03, 10:45 AM 

I ■ I I I 


Collin-Osdoby et al. 


Fig. 1. Chick OC immunomagnetic purification and bone pit resorption. (A) Chick 
OC from 35% Percoll preparations were affinity captured and purified via their bind- 
ing to MAb 121F-coupled magnetic beads. Phase-contrast microscopy reveals numer- 
ous beads avidly attached to and covering OC cell surfaces. (B) SEM appearance of 
6% Percoll purified chick OC preparations cultured on plastic. Scale bar = 50 [im. (C) 
SEM analysis of MAb 121F immunomagnetically isolated chick OC and an associated 
resorption pit formed during culture on bone. Scale bar = 10 [im. 

chick OCs appear as large multinucleated cells of varying sizes and shapes, 
which have a grainy cast and frequently one or more pseudopodial extensions 
per cell. Immunomagnetically isolated chick OCs are typically decorated with 
multiple beads per cell and can be so thoroughly coated with MAb-conjugated 
beads that they resemble a ball of beads (Fig. 1A). On culture of 
immunomagnetically captured OCs on bone, ivory, glass, or plastic, the cells 
spread out and internalize the beads, rather than shedding them as do 
nonphagocytic cells. 



2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 77 

OC morphology and ultrastructure can be analyzed in greater detail using 
transmission (TEM) or scanning (SEM) electron microscopy as detailed in 
other chapters of this volume. Features that are characteristic of OCs and should 
be evident by TEM include: multiple nuclei often clustered within the cell and 
varying in number between cells, abundant mitochondria, numerous vesicles, 
extensive vacuolation, well-developed perinuclear Golgi complexes, promi- 
nent rough endoplasmic reticulum, free polysomes, and ruffled border mem- 
brane and clear zone domains. By SEM, chick OCs cultured on plastic typically 
appear as large cells having a complex morphology with many fine filopodial 
projections, microvilli, and membrane blebs visible over the cell surface, and a 
peripheral cytoplasmic skirt (Fig. IB). When cultured on bone or ivory, chick 
OCs appear by SEM either as large domed cells actively engaged in excavating 
a resorption cavity or as stretched inactive cells having a motile phenotype and 
characteristic leading and trailing membrane domains (Fig. 2F). OCs associ- 
ated with resorption pits often exhibit membrane projections stretched back 
over a portion of the well-excavated lacuna that has exposed collagen fibrils 
(Fig. 2F). Resorption pits formed by cultured chick OCs are typified by 
multilobulated excavations or long resorption tracks (which also may be 
multilobulated) or, less often, as a unilobular cavity adjacent to or underlying 
an OC actively involved in resorption (Fig. 2F). Immunomagnetically isolated 
OCs may exhibit less pit resorbing activity (Fig. 1C). The morphology of re- 
sorption pit excavations can be examined more thoroughly by SEM following 
the removal of OCs from the bone or ivory substrate (either initially before a 
gold coating step or after viewing the sample and then recoating with gold to 
visualize the pits alone). 

3.3.2. Cytochemical Staining 

A quick and easy method for discriminating nuclear and cytoplasmic detail 
in cultured chick OCs is by use of the general differential stain Difquik, which 
is simply incubated with the fixed chick OCs for several min and then rinsed 
off. However, by far the most commonly used stain to visualize OCs is based 
on their high level of TRAP activity, which is upregulated early in OC devel- 
opment and is essential for their resorption of bone. Although TRAP activity is 
not completely specific for OCs alone, this cytochemical stain readily identi- 
fies OCs in bone tissue sections and in isolated OC preparations (Fig. 3A,C). 
In addition to detecting the relative levels of TRAP activity associated with 
individual OCs by staining, cell extracts can be quantitatively evaluated for 
TRAP activity levels using a microplate enzymatic assay and such activity 
normalized for cell extract protein (12). 

To stain for TRAP activity using freshly prepared staining solutions (see 
Subheading 2.3. and Note 2): 

06/Collin/65-88/F1 77 I 2/26/03, 10:45 AM 

I ■ I I I 


Collin-Osdoby et al. 


Fig. 2. Chick OCs form resorption pits on ivory and phagokinetic tracks on gold- 
coated coverslips. (A-C) 6% Percoll-purified chick OCs were cultured on ivory (2-3 d) 
and harvested for TRAP staining and resorption pit analysis. As viewed by light mi- 
croscopy, OCs formed multilobulated resorption tracks that frequently were com- 
posed of connecting resorption lacunae. These represent periods of OC attachment 



2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 79 

1. Remove the conditioned medium from the cells (and either discard or save for 
other analyses). 

2. Rinse cells quickly three times with warm HBSS (tilt dish to add and remove 
solutions gently). 

3. Add 1% PF-HBSS solution (-300 fxL per 48- well, 500 \iL per 24- well-plate) and 
fix cells for 15 min. at room temperature. 

4. Remove the fixative and rinse three times with HBSS and once with deionized water. 

5. Air-dry the samples overnight (to permeabilize the cells) or, alternatively, incu- 
bate in -20°C methanol for several min, followed by a rinse with water (there is 
no need then to air-dry). 

6. Add enough staining solution to cover cells in the wells and incubate the dish or 
plate at 37°C for 1 h in the dark. 

7. Remove the stain and rinse the samples several times with water. 

8. Air-dry the samples (on the dish or ivory) or mount coverslips using fine twee- 
zers to pick up a coverslip and invert it onto a drop of glycerol buffered-mount- 
ing medium spotted onto a glass slide (these should then be stored at 4°C and 
rewarmed before viewing in a microscope). 

Alternatively, a commercially available staining kit (Sigma cat. no. 386) 
can be used as directed by the manufacturer (see Note 2). 

3.3.3. Antigenic Profile 

Together with specific morphological features and high TRAP activity lev- 
els, chick OCs exhibit various characteristic surface markers that are commonly 
monitored, as they are not expressed (or only at low levels) by related mono- 
cytes, macrophages, or macrophage polykaryons. These include expression of 
av(33 integrin (vitronectin receptor), H + -ATPase proton pump, carbonic anhy- 
drase II, and a series of antigens recognized by anti-OC MAbs including the 
12 IF MAb-reactive OC-specific membrane glycoprotein (Fig. 3B,D,F,G). 
Most, if not all, of these OC markers play an essential or important role in the 
bone resorptive function and/or survival of OCs. Using specific antibodies, 

Fig. 2. (continued) and pit formation, followed by OC movement to an adjacent area 
of ivory for further resorption. (D) Resorption pits viewed by darkfield reflective light 
microscopy, as performed for quantifying the number and areas of resorption pits 
within the exact same fields evaluated for OC numbers. (E) 6% Percoll-purified chick 
OCs cultured on gold coated coverslips for 16 h and subsequently stained for TRAP 
activity. OCs phagocytose the gold particles and thereby clear a path during their 
movement across the gold-coated coverslip. The numbers and areas of such 
phagokinetic tracks are measured and expressed relative to the numbers of associated 
OCs. (F) SEM analysis of 6% Percoll-purified chick OCs engaged in bone pit resorp- 
tion on ivory. Note the deep, well excavated lacunae that are typically formed by chick 
OCs. Scale bar = 50 [im. 

06/Collin/65-88/F1 79 I 2/26/03, 10:45 AM 

I ■ I I I 


Collin-Osdoby et al. 


Fig. 3. Cytochemical and immunostaining analysis of 6% Percoll purified-chick 
OCs. (A,C) Chick OCs cultured on plastic (1-2 d) and stained for TRAP activity. 
(B,D) Chick OCs cultured on plastic and immunostained using a MAb to the 
vitronectin receptor, integrin av(33 (LM 609) and a biotin-streptavidin |3-galactosi- 
dase detection system. (E) Chick OCs cultured on plastic, fixed and permeabilized 
(Triton X-100), and double stained with rhodamine phalloidin to label cytoskeletal F- 



2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 81 

these markers can be detected on the surface of chick OCs cultured and fixed 
on bone, ivory, glass, or plastic via immunostaining, alone or in combination 
with TRAP staining, F-actin cytoskeletal staining (with rhodamine-labeled 
phalloidin), and/or DAPI nuclear labeling (Fig. 3A-G). The relative surface 
level expression of these protein markers can also be measured for chick OCs 
following their fixation (as whole cells) in 96-well microtiter dishes and quan- 
titative analysis by ELISA as detailed elsewhere (12). Alternatively, these 
markers may be monitored in total cell extracts or membrane lysates of chick 
OCs by ELISA, gel electrophoresis (with or without immunoprecipitation), or 
immunoblotting. OCs also express high intracellular levels of pp60 csrc , a criti- 
cal signal molecule required for OC bone resorption. This cytoskeletally asso- 
ciated protein can be detected by immunostaining in permeabilized cells (e.g., 
incubate fixed OCs in 0.1% Triton X-100 for 30 min prior to blocking) or by 
immunoblotting of electrophoresed cell extracts, with or without probing for 
phosphorylation status. A general protocol for immunostaining is given here 
(use a minimum of 250 uL of reagent per well of a 24 well-plate): 

1. After culture of OCs on coverslips, bone, or ivory, rinse the tissue culture wells 
gently and fix as for TRAP staining above (see Subheading 3.3.2., steps 1-4). 

2. Do NOT let the samples dry, but, instead, immediately process them for antibody staining. 

3. Block nonspecific protein binding sites by incubating with blocking solution for 
1 h at room temperature. 

4. Incubate with appropriate dilutions of antibodies directed against OC antigens 
for 1 h at room temperature. Incubate one set of samples (representative of each 
test condition) in blocking solution alone (in place of antibody) to serve as nega- 
tive controls for nonspecific background staining. 

5. Rinse three times briefly and once for 10 min. with PBS. 

6. Incubate with a secondary biotin-conjugated antibody directed against the pri- 
mary antibody for 1 h at room temperature. 

7. Rinse three times briefly and once for 10 min. with PBS. 

8. Incubate for 30 min in the dark with streptavidin conjugated with FITC (or Texas red). 

9. Rinse three times briefly and once for 10 min. with PBS. 

10. Mount specimens onto glass slides with glycerol-buffered mounting medium {see 
Subheading 3.2.2., step 8), store in the dark at 4°C, and rewarm slides before 
viewing in a microscope. 

Fig. 3. (continued) actin (red) as well as DAPI to label multiple nuclei within OCs 
(blue). Note the peripheral actin ring formation characteristic of mature OCs. (F,G) 
Chick OCs cultured on plastic and immunostained with MAb 12 IF using a biotin- 
streptavidin FITC detection system. Both av|33 and the 121F antigen become highly 
expressed on the OC cell surface during their differentiation into bone-resorptive 
multinucleated cells, and each of these OC markers plays an important functional role 
in the resorption of bone by OCs. 

06/Collin/65-88/F1 81 1 2/26/03, 10:45 AM 

I ■ I I I 


82 Collin-Osdoby et al. 

11. If desired, OCs on coverslips can be briefly reacted with a membrane-permeable 
fluorescent dye to label the nuclei (bright blue) by incubation in a 1:300 dilution 
of DAPI in HBSS for 1 min, followed by two rinses in HBSS before mounting. 
Immunostained samples can also be stained for TRAP activity (after step 9 above) 
before being mounted on glass slides. 

Antigen detection on OCs adherent to bone or ivory is difficult to measure 
using a fluorescent system unless confocal microscopy is used (see the chapter 
by Nesbit and Horton, this volume). A histochemical detection method is more 
appropriate for this purpose, and this method also works well to immunostain 
OCs cultured on glass or plastic in place of the fluorescent system. We prefer 
to use antibodies coupled to (3-galactosidase (which has negligible background 
problems and requires no specific blocking), but other enzymes (e.g., horse- 
radish peroxidase) can also be used with good results if endogenous enzyme 
activities are quenched (see the chapter by Bord, this volume). 

For (3-galactosidase-based immunostaining, perform steps 1-7 of the proto- 
col given earlier in this subheading. Then continue as follows: 

1 . Incubate for 30 min with streptavidin conjugated with |3-galactosidase in buffer A. 

2. Rinse five times with buffer A over 30 min. 

3. Incubate for 30 min (or longer, if necessary) in the dark with substrate solution. 

4. Rinse five times over 30 min with PBS. 

5. Store bone or ivory slices dry before viewing. (If using this protocol for cells on 
coverslips, do not dry but, instead, mount the coverslips onto glass slides as 
described in step 10 earlier in this subheading). 

6. OCs that have been immunostained by this method (on bone, ivory, glass or plas- 
tic) can also be subsequently double stained for TRAP activity (see Subheading 

3.3.4. Molecular Profile 

Both Percoll-fractionated and immunomagnetically purified chick OCs can 
serve as sources of RNA for analyzing the relative gene expression levels of 
various OC phenotypic and functional markers either in freshly isolated cells 
or following OC culture on plastic, bone, or ivory in the presence or absence of 
modulators. Although it can be difficult to obtain sufficient RNA to examine 
changes in mRNA expression in response to varying conditions by northern 
analysis, OC preparations readily provide sufficient RNA for more sensitive 
ribonuclease protection assay (RPA) or reverse transcription-polymerase chain 
reaction (RT-PCR) applications (13). Typically, one well of a 24-well dish 
seeded with 200,000 viable chick OCs (in 250 uL of medium) yields 2-5 ug of 
total RNA, of which 1 00 ng-1 ug may be used per RT-PCR reaction, or up to 
5 fxg for a single RPA assay. Currently, chicken-specific primers for PCR 
amplification may be available only for a limited subset of chick OC markers, 

06/Collin/65-88/F1 82 1 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 83 

however, it is sometimes possible to generate others as needed based on prim- 
ers or sequences reported for OC genes cloned from human or mouse or other 
species. Thus, regions corresponding to high interspecies homology and gene 
specificity can be chosen for initial PCR attempts with chick OCs; any ampli- 
fied products are then sequenced and optimal chicken-specific primers can then 
be designed based on the sequences obtained. Similarly, species homologous 
primers can be used to amplify chicken-specific genes and the PCR products 
cloned into appropriate vectors for the preparation of probes for RPA analyses of 
chicken OCs (13). Methods for RNA preparation are given elsewhere in this 
volume (see also chapter by Stewart and Mann, this volume). 

3.3.5. Motility (see Note 4) 

Because OCs are very large and vary considerably in size (and number of 
nuclei), it can be difficult to perform classical chemotaxis/chemokinesis 
experiments through porous membranes to measure OC movement in response 
to various agents. However, OC movement can easily be monitored after cul- 
ture on gold-coated coverslips, because OCs phagocytose the gold and thereby 
generate a cleared track in their wake (10). To perform this assay, 6% Percoll- 
purified OCs are used (see Subheading 3.1.1.). 

1. Resuspend the cells gently in 6-8 mL of culture medium. 

2. Plate out 0.5 mL (-100,000 OCs) per well of a 24-well dish containing rinsed and 
prewetted gold coated coverslips (see Subheading 2.6.). 

3. Culture the cells for 2-3 h to allow OCs to attach. 

4. Remove the nonadherent cells and add fresh medium with or without modulators. 

5. Culture the cells for 16-24 h. 

6. Rinse gently three times with warm HBSS. 

7. Fix the cells with 1% PF-HBSS for 15 min. at room temperature. 

8. Rinse as in step 6. 

9. Stain the cells for TRAP activity (see Subheading 3.3.2. and Note 2). 

10. Determine the number of TRAP-stained OCs, the number of phagokinetic tracks, 
and the cleared area of each track within a constant number of random adjacent 
fields using a microscope fitted with an ocular reticle and computer linked to an 
image analysis system (Fig. 2E). Calculate the mean track area and the total area 
of gold cleared. Also present the data as normalized values representing the num- 
ber of tracks per OC and the mean area cleared per OC. 

3.3.6. Bone Resorption 

OCs and other phagocytic cells can all resorb (ingest and degrade) small 
particles of bone in vitro or secrete acid that dissolves hydroxyapatite (e.g., on 
calcium phosphate coated dishes). However, only OCs are capable of excavat- 
ing resorption pit lacunae on bone. This is therefore the key defining attribute 
and best assay for evaluating the bone resorptive function of fully developed 

06/Collin/65-88/F1 83 1 2/26/03, 10:45 AM 

I ■ I I I 


84 Collin-Osdoby et al. 

OCs. Because the number of new sites initiated and the rate of resorption by 
OCs are major parameters controlling bone remodeling in both normal and 
pathological states, the in vitro bone pit resorption assay has therefore become 
a very valuable investigational tool. The data obtained from this analysis 
reveals information about whether a modulator has altered the number of OCs 
on the bone or ivory (possibly reflecting effects on integrin-mediated attach- 
ment, cell survival, or development), the number of pits formed (reflecting the 
activation of OCs for initiating pit resorption), and the area of bone or ivory 
resorbed (overall or per pit, reflecting the amount and rate of resorption by OCs). 

1. Culture chick OCs (6% Percoll purified) on bone or ivory in the presence or 
absence of modulators for 30-40 h (see Note 13). 

2. Rinse, fix, and stain for TRAP activity as described in Subheading 3.3.2. 

3. Evaluate resorption by using a microscope fitted with an ocular reticle and com- 
puter linked to an image analysis system (see Chapter 1 1 by van 't Hof et al. for 
a detailed description of a system) (Fig. 2A-C). 

4. Count the number of TRAP-stained OCs within a constant number of random 
fields per bone slice, usually measured consecutively from an arbitrary starting 
location on the edge of the chip that is marked with a dot using a permanent ink 
marker. The number of fields chosen for analysis should encompass at least 100- 
300 OCs per bone chip (typically -20 fields or half of the chip). To ensure that 
the exact same fields will subsequently be analyzed for resorption pits, mark or 
draw the fields that have been evaluated for each chip on a grid log sheet. 

5. After all the bone chips have been analyzed for OC numbers (to ensure that no 
category has too few OCs), remove the OCs from the bone surface by soaking the 
chip for 1 min in 0.2 M NH 4 OH, rubbing the entire surface with gloved fingers, 
repeating this treatment a second time, and then rinsing the chip in deionized water. 

6. Quantify the number and planar area of each resorption pit contained within the 
fields evaluated for OC numbers in step 4 using darkfield reflective light 
microsopy (Fig. 2D). 

7. Express resorption measures as the mean number of OCs, number of pits, and 
total areas resorbed in this constant number of fields for each experimental con- 
dition. Also normalize the data to yield the mean number of pits per OC, area 
resorbed per OC, and area (size) per individual pit. 

8. In general, several trials, with four to six replicates each for control and treated 
groups, should be performed to achieve statistically significant results. 

4. Notes 

1. OC-specific antibodies that can be used for this purpose include MAb 121F 
(available upon request from our laboratory) or antibodies to integrin av(33 (e.g., 
23C6 or LM609; see the chapter by Coxon et al., this volume). 

2. Cytochemical staining for TRAP activity is routinely performed using either 
reagents freshly prepared in the laboratory or a commercially available staining 
kit (Sigma cat. no. 386) as directed. Because the intensity of TRAP staining tends 

06/Collin/65-88/F1 84 I 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 85 

to be stronger with the freshly prepared reagents, this protocol is preferred for 
tracking changes in OC TRAP activity (during development or in response to 
modulators), for discriminating OCs in quantitative resorption pit analyses, and 
for double staining of immunostained OCs. 

3. Secondary antibodies may be directly conjugated with an enzymatic or fluores- 
cent probe, but greater sensitivity is achieved if a biotinylated secondary anti- 
body is used to amplify the primary antibody signal (e.g., biotinylated goat 
anti-mouse IgG for detecting a primary mouse MAb to an OC antigen). 

4. We find that gelatin precoating enables the gold particles to remain more reliably 
adherent to the glass coverslips. BSA cannot substitute for gelatin, as the former 
interferes with chick OC movement on the gold-coated coverslips. It is important 
to avoid plating too many OCs on the gold-coated coverslips because the result- 
ant overlapping tracks become difficult to analyze. Similarly, incubation times 
with chick OCs should not exceed 16-24 h generally because the phagokinetic 
tracks may become too long and convoluted (and therefore overlap significantly), 
migration differences in response to agents may become less apparent, and OCs 
may cease movement when overloaded with ingested gold particles. 

5. Young posthatch growing chicks represent a highly abundant source of OCs, 
whose numbers are further increased by maintaining the chicks on a low-calcium 
feed diet. However, it is important that the level of calcium in this feed does not 
fall below 0. 15% or the bones of the young chicks will become so soft and weak 
that the birds are unable to stand to walk, eat, or drink. The feed is prepared by 
special order (Purina) and can be stored at 4°C (kept dry to avoid mold) for up to 
12 mo. In addition, the birds should have free access to tap water (not deionized 
water which makes them too weak when they are on the low calcium feed). As 
opposed to the millions of OCs that are obtained from chicks, more than 1000- 
fold fewer OCs are typically isolated from mouse, rat, or rabbit bone prepara- 
tions, and only negligible OC numbers from most human bone tissue. Besides 
sharing all the morphological, phenotypic, and antigenic properties as well as most 
or all of the regulatory responses observed with OCs from other species, chick OCs 
are the most highly active and aggressive OC species for bone pit resorption. Con- 
sequently, chick OCs provide a particularly sensitive assay system to measure the 
regulatory influences of various agents (many of which typically act to suppress, 
rather than stimulate, OC activity) on OC-mediated bone resorption. 

With the recent discovery of the RANKL-RANK-OPG regulatory pathway control- 
ling OC development, resorption, and survival, some of the restrictions on studies 
with OCs from other species have been alleviated as it is now possible to generate 
OCs in vitro from OC precursors present in primary cell preparations (e.g., avian, 
mouse, or human bone marrow or circulating monocytes) or cell lines (murine RAW 
264.7 cells). Details of such procedures are given in the chapters by Sabokbar and 
Athanasou, Flanagan amd Massey, and Collin-Osdoby et al. (Chapter 12). 
Despite such advances, it will often still be necessary to compare the responses 
obtained with such in vitro generated OCs against those observed with isolated 
OCs formed in vivo for each agent under investigation. 

06/Collin/65-88/F1 85 1 2/26/03, 10:45 AM 

I ■ I I I 


86 Collin-Osdoby et al. 

6. OC viability is dependent on the total length of time that it takes from the removal 
and processing of the bones until the 6% Percoll fractionated or 
immunomagnetically isolated OCs are placed into culture. It is best if this time 
does not exceed 6-7 h, as each additional hour will negatively impact on the final 
OC viability obtained. If the bone marrow is to be used (e.g., for OC precursor 
studies) from these same chicks, one person should be designated to blow out the 
bone marrow in a sterile hood from a group of bones while other individuals are 
harvesting or cleaning the remaining bones, and the dishes should be passed back 
and forth until they are all completed. One person should then continue with the 
bone marrow cell preparation independently from those working on isolating OCs 
from the marrow stripped bones (12). 

7. Use only wide-bore pipets or tips for any work in isolating or manipulating OCs to 
avoid fragmenting these large multinucleated cells. Also, care should be taken to 
resuspend, mix, or vortex OC preparations gently and for as little time as necessary. 

8. Immunomagnetic isolation of OCs provides a rapid and highly efficient way of 
purifying OCs from mixed cell populations. The magnetic beads coupled with 
121F MAb are not species restricted and can therefore also effectively purify 
OCs from rat, rabbit, and human sources. Similarly, other anti-OC MAbs can 
also be coupled to such magnetic beads and used to purify OCs via this procedure 
(see also Note 1). The procedure is performed on ice to prevent ingestion of the 
beads by OCs. However, because MAb 12 IF (bivalent or Fab fragments) par- 
tially inhibits OC bone pit resorption (14), and immunomagnetic yields of OCs 
are lower than with Percoll purifications, immunomagnetically isolated OCs are 
therefore considered most useful for: (1) obtaining highly purified OC samples 
for molecular or biochemical analysis and (2) confirming that responses observed 
in Percoll-purified OC preparations can be attributed to OC-specific effects. 

9. For optimum OC immunomagnetic capture, it is important to avoid stirring the 
MAb-coupled beads with the cells either too fast (which interferes with their 
attachment) or too slow (which reduces their binding due to poor mixing). Best 
results are achieved with 35% Percoll-separated OC preparations used as the start- 
ing material as opposed to more crude preparations because the latter yields less 
pure OC populations and matrix reassembly is more problematic. Six percent 
Percoll preparations typically yield fewer immunomagnetically isolated OCs. 

10. As an alternative to seeding OCs onto individual bone slices, OCs may be seeded 
in 2.5 mL of medium onto -24 bone or ivory discs spread out to cover fully the 
bottom of a 35-mm dish. After allowing OCs to attach selectively to the bone or 
ivory pieces for 2.5-3 h, the nonadherent cells are removed, adherent OCs are 
gently washed with medium, and the discs are individually lifted out with sterile 
tweezers. Each disc is then placed into one well of a 48-well culture dish with 
250 [ih of fresh medium. Modulators are typically administered (diluted into 50 |xL 
of medium) at this time (to yield a final volume of 300 jxL in each well) and the 
cells are cultured for 30-40 h before harvest. 

11. For resorption, OCs are routinely cultured for 30-40 h before harvest. For his- 
tochemical, enzymatic, or immunocytochemical analysis, OCs are cultured for 1 

06/Collin/65-88/F1 86 1 2/26/03, 10:45 AM 

I ■ I I I 


Chicken Osteoclast Cultures 87 

or 2 d before analysis. For molecular studies, RNA may be harvested directly 
from the Percoll-fractionated or immunomagnetically captured cells. Alterna- 
tively, RNA can be harvested from OCs cultured on bone, ivory, or plastic for up 
to 3 d in the presence or absence of modulators. OC survival is enhanced if the 
cells are cultured on bone or ivory (due to integrin-mediated survival signals) as 
opposed to glass or plastic, so experiments under the latter conditions should be 
limited to a few days at most. Although some studies have indicated that OCs 
resorb better under slightly acidic conditions, we find that chick (and human) OC 
performance is actually better in a-MEM supplemented with 5% FBS. 

12. Immunomagnetically purified OCs can be cultured and will form resorption pits 
on bone or ivory, but because they will ingest beads rapidly at 37°C (rather than 
shed the beads during culture), it is important to consider that the antibody and/or 
bead engagement of the OCs cell surface could affect their physiology or resorp- 
tive function (Fig. 1A,C). It is possible to remove many, but not all, of the beads 
from the outer surface by physical (strong vortex-mixing) or biochemical (low 
pH, protease digestion) methods, performed at 4°C, although some cell damage 
may occur during these procedures (9). 

13. Because ethanol inhibits OC bone resorption, it is very important that the bone or 
ivory slices are well rinsed (and soaked for > 3 h) in HBSS before the cells are 
plated onto them. In general, resorption pit analysis using bone is somewhat more 
complicated than with ivory owing to the need to distinguish Haversian canals in 
the bone apart from the pits made by the cultured OCs. In our replicate studies 
using bone and ivory, no substrate-dependent differences have been noted to date 
in either basal or modulator evoked resorption parameters for isolated chicken 
OCs. Therefore, although ivory can be more difficult to obtain than bovine bone, 
it is preferable to use for quantitative resorption pit analysis. 

Acknowledg merits 

This work was supported by NIH Grants AR32927, AG 15435, and AR427 1 5 
to P. O. 


1. Hall, T. and Chambers, T. (1996) Molecular aspects of osteoclast function. 
Inflctmm. Res. 45, 1—9. 

2. Roodman, G. (1996) Advances in bone biology — the osteoclast. Endoc. Rev. 17, 

3. Suda, T., Udagawa, N., Nakamura, I., Miyaura, C, and Takahashi, N. (1995) 
Modulation of osteoclast differentiation by local factors. Bone 17, S87-S91. 

4. Kanis, J. (1995) Bone and cancer: pathophysiology and treatment of metastases. 
Bone 17, 101S-105S. 

5. Mundy, G. (1993) Cytokines and growth factors in the regulation of bone remod- 
eling. /. Bone Miner. Res. 8, S505-S510. 

6. Wiebe, S., Hafezi, M., Sandhu, H., Sims, S., and Dixon, S. (1996) Osteoclast 
activation in inflammatory periodontal diseases. Oral Dis. 2, 167-180. 

06/Collin/65-88/F1 87 I 2/26/03, 10:45 AM 

I ■ I I I 


88 Collin-Osdoby et al. 

7. Manolagas, S., Bellido T., and Jilka, R. (1995) New insights into the cellular, 
biochemical, and molecular basis of postmenopausal and senile osteoporosis: roles 
of IL-6 and gpl30. Int. J. Immunopharmacol. 17, 109-116. 

8. Oursler, M., Collin-Osdoby, P., Anderson, F., Li, L., Webber, D., and Osdoby, P. 
(1991) Isolation of avian osteoclasts: improved techniques to preferentially purify 
viable cells. /. Bone Miner. Res. 6, 375-385. 

9. Collin-Osdoby, P., Oursler, M., Webber, D., and Osdoby, P. (1991) Osteoclast- 
specific monoclonal antibodies coupled to magnetic beads provide a rapid and 
efficient method of purifying avian osteoclasts. /. Bone Miner. Res. 6, 1353-1365. 

10. Minkin, C. (1982) Bone acid phosphatase: tartrate-resistant acid phosphatase as a 
marker of osteoclast function. Calcif. Tissue. Int. 34, 285-290. 

11. Owens, J. and Chambers, T. (1993) Macrophage colony-stimulating factor (M- 
CSF) induces migration in osteoclasts in vitro. Biochem. Biophys. Res. Commun. 
195, 1401-1407. 

12. Collin-Osdoby, P., Oursler, M., Rothe, L., Webber, D., Anderson, F., and Osdoby, 
P. (1995) Osteoclast 121F antigen expression during osteoblast conditioned 
medium induction of osteoclast-like cells in vitro: relationship to calcitonin 
responsiveness, tartrate resistant acid phosphatase levels, and bone resorptive 
activity. /. Bone Miner. Res. 10, 45-58. 

13. Sunyer, T., Rothe, L., Kirsch, D., et al. (1997) Ca 2+ or phorbol ester but not 
inflammatory stimuli elevate inducible nitric oxide synthase messenger ribo- 
nucleic acid and nitric oxide (NO) release in avian osteoclasts: autocrine NO 
mediates Ca 2+ -inhibited bone resorption. Endocrinology 138, 2148-2162. 

14. Collin-Osdoby, P., Li, L., Rothe, L., et al. (1998) Inhibition of osteoclast bone 
resorption by monoclonal antibody 121F: a mechanism involving the osteoclast 
free radical system. /. Bone Miner. Res. 13, 67-78. 

06/Collin/65-88/F1 88 1 2/26/03, 10:45 AM 

I ■ I I I 


Isolation and Purification of Rabbit Osteoclasts 

Fraser P. Coxon, Julie C. Frith, Helena L. Benford, 
and Michael J. Rogers 

1. Introduction 

Osteoclasts are notoriously hard to study because of the difficulty in obtain- 
ing pure populations of cells in large numbers for biochemical and molecular 
analyses. Unlike with other rodents (e.g., mice and rats), mature osteoclasts 
can be obtained from rabbits in relatively large numbers and can be purified 
easily. For some studies, such primary cultures of authentic osteoclasts may be 
preferable to osteoclast-like cells generated in vitro from bone marrow cul- 
tures. Isolated rabbit osteoclasts are capable of resorbing mineralized substrates 
in vitro and are therefore useful for assessing the effect of pharmacologic agents 
on osteoclast-mediated bone resorption (1,2). We and others have also 
extracted protein or RNA from purified rabbit osteoclasts for studies on meta- 
bolic processes in osteoclasts, or molecular studies on osteoclast biology using 
Western blotting, enzyme assays or reverse transcriptase-polymerase chain 
reaction (RT-PCR) (3-7). 

We routinely isolate osteoclasts from the long bones of neonatal rabbits 
using a method adapted from that of Tezuka et al. (8) (see Subheading 3.1.). 
Isolated rabbit osteoclasts can then be cultured on plastic, glass, or mineralized 
substrates (such as bovine cortical bone, elephant ivory, or whale dentine). 
Culturing osteoclasts on glass coverslips (in multiwell plates) is useful for 
immunocytochemistry, as the coverslips can be mounted onto glass slides, 
enabling cells to be visualized using an upright microscope. 

For some applications, such as preparation of osteoclast lysates for Western 
blot analysis, osteoclasts must be further purified from contaminating bone 
marrow cells. This can be done either by further washing culture dishes with 
phosphate-buffered saline (PBS), or by removing contaminating adherent cells 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


07/Rogers/89-100/F1 89 1 2/26/03, 10:45 AM 

I ■ I I I 


Coxon et al. 


Fig. 1 . (A) Phase-contrast photograph of multinucleated, rabbit osteoclasts cultured 
in a plastic Petri dish following purification with PBS. Scale bar = 25 [im. (B) A rabbit 
osteoclast showing VNR expression by fluorescence immunostaining. 


V <Jp ~ 

i *jd^ 



lul B 

Fig. 2. Phase-contrast photographs of a rabbit bone marrow culture (A) after 7 d, 
showing a developing multinucleated, osteoclast-like cell (arrow) beneath the stromal 
cell layer; (B) following purification of osteoclast-like cells after 10 d (multinucleated 
cell shown in inset). Scale bars = 20 (xm. 

using a solution of pronase-EDTA (see Subheading 3.2.)- This provides cul- 
tures of >95% pure, tartrate-resistant acid phosphatase (TRAP)-positive, multi- 
nucleated osteoclasts and mononuclear, prefusion osteoclasts (Fig. 1A). 

When isolating mature osteoclasts we typically achieve a yield of approx 
5 x 10 4 purified osteoclasts from each rabbit. It is also possible to generate 
larger numbers of osteoclast-like cells (TRAP -positive multinucleated cells, 
capable of resorption) without having to euthanize more rabbits. Indeed, from 
each rabbit it is possible to generate up to 16 semiconfluent 10-cm Petri dishes 
of osteoclast-like cells (Fig. 2). This is achieved by culturing the nonadherent 
cells in the presence of 1,25-dihydroxyvitamin D 3 [l,25-(OH) 2 D 3 ] over a 



2/26/03, 10:45 AM 

I ■ I I I 

Rabbit Osteoclasts 



- J 1 

Fig. 3. Scanning electron micrograph of an immunomagnetically isolated rabbit 
osteoclast cultured on ivory. The osteoclast is associated with a resorption lacuna and 
still has magnetic beads attached. Scale bar = 10 p,m. 

period of 10 d, using a method modified from David et al. (9) (see Subhead- 
ing 3.3.) 

For in vitro applications, purification of osteoclasts on culture dishes by 
pronase-EDTA digestion is sufficient to provide pure osteoclasts for biochemi- 
cal studies. However, a pure population of osteoclasts can also be isolated 
directly from rabbit bones, without prior cell culture in vitro. This is particu- 
larly useful when studying the effects of pharmacological agents on osteoclasts 
in vivo. We have modified a technique (see Subheading 3.4.) developed by 
Collin-Osdoby and colleagues (10), that involves separation of osteoclasts from 
a mixed cell suspension using immunomagnetic beads and the 23C6 monoclonal 
antibody (Fig 3). The latter specifically recognizes the a v (3 3 /vitronectin receptor 
(VNR) integrin, which is highly expressed on osteoclasts (11) (Fig. IB). 

Rabbit osteoclasts in culture can be identified using markers that are highly 
abundant in these cells, for example, by staining for TRAP (see Subheading 
3.5.1.) and immunological detection of the vitronectin receptor/a v (3 3 integrin 
(Fig. 1A) (see Subheading 3.5.2.). Enzyme histochemical studies have shown 
that osteoclasts contain TRAP in abundance (12); therefore, staining for activity 
of this enzyme is useful for enabling the number of osteoclasts in culture to be 
counted, particularly when the cells are cultured on a substrate, such as ivory, 


2/26/03, 10:45 AM 

I ■ I I I 


Coxon et al. 



-c..; «■ 

1 : H 






*> * 

v ; 

Fig. 4. Resorptive activity of rabbit osteoclasts cultured on ivory discs. (A) Scan- 
ning electron micrograph showing a cultured rabbit osteoclast (arrow) adjacent to a 
resorption pit (asterisk). Scale Bar = 10 Jim. (B) A rabbit osteoclast stained with 
TRITC-phalloidin, exhibiting a characteristic F-actin ring. (C) Reflected light photo- 
graph of resorption pits excavated by rabbit osteoclasts. (Reduced from original mag- 
nification, xlO). 

on which cells cannot easily be seen by light microscopy. It should be noted, 
however, that this enzyme is not specific for osteoclasts and is present in other 
cell types, such as alveolar macrophages. To stain cells for TRAP activity we 
use a method adapted from that of van't Hof et al. (13). When rabbit osteo- 
clasts are cultures on a mineralized substrate such as ivory, actively resorbing 
osteoclasts can be identified by the presence of a characteristic ring of F-actin 
(see Subheading 3.5.3.) or by their ability of the osteoclasts to excavate 
resorption pits in the substrate (see Subheading 3.5.4.) (Fig. 4). 

2. Materials 

2.1. General Reagents 

1. a-Minimum essential medium (a-MEM) supplemented with 100 U/mL of peni- 
cillin, 100 |xg/mL streptomycin, and 1 mM glutamine. 

2. Fetal calf serum (FCS). 

3. PBS. 

4. 4% Formaldehyde in PBS. 

2.2 Isolation and Purification of Rabbit Osteoclasts 

1. Sharp scissors. 

2. Blunt-ended forceps. 

3. Disposable scalpel (for removing tissue). 

4. Autoclaved scalpel handle and disposable scalpel blade (for mincing bones). 

5. 10-cm diameter glass Petri dishes. 

6. PBS containing 0.001% (w/v) pronase, 0.002% (w/v) EDTA. Filter-sterilize (0.2-^m 
filter) before use. Pronase can be prepared as a concentrated stock solution in PBS 
and stored frozen as aliquots at -20°C, then diluted in PBS-EDTA before use. 



2/26/03, 10:45 AM 

I ■ I I I 


Rabbit Osteoclasts 93 

2.3. Generation of Rabbit Osteoclast-Like Cells 

1. l,25-(OH) 2 D 3 (Sigma, Poole, UK) 

2.4. Isolation of Rabbit Osteoclasts Using Immunomagnetic Beads 

1. 23C6 (Anti-a v |3 3 ) monoclonal antibody (Mab) (Serotec, Oxford, UK). 

2. 0.1% (w/v) Bovine serum albumin (BSA) in PBS. 

3. Magnetic beads conjugated to rat anti-mouse IgG (e.g., Dynal, Lake Success, NY). 

4. Magnet or magnetic particle concentrator. 

2.5. Staining for TRAP 

1. 10 mg/mL of Naphthol-AS-BI-phosphate substrate in dimethyl formamide 
(stable at 4°C for about 2 wk). 

2. 4% (w/v) Sodium nitrite. 

3. Pararosanilin: Add 1 g of pararosanilin to 20 mL of dH 2 then add 5 mL concen- 
trated HC1. In a fume hood, heat the solution carefully with constant stirring in a 
water bath for 30 min, then filter after cooling (stable at 4°C for several months). 

4. Veronal buffer: 1 1.7 g/L of anhydrous sodium acetate, 29.4 g/L of veronal (bar- 
bital) in dH 2 (toxic solution) (stable at 4°C for several months). 

5. 0.1 N acetate buffer, pH 5.2: dissolve 0.82g of anhydrous sodium acetate in 
100 mL of dH 2 0. Adjust the pH of this solution to 5.2 using a solution of 
0.6 mL of glacial acetic acid made up to 100 mL with dH 2 (stable at 4°C 
for several mon). 

6. Acetate buffer plus tartrate: Add 2.3 g of sodium tartrate to 100 mL of acetate 
buffer, to give a stock solution of 100 mM tartrate (stable at 4°C for several mon). 

2.6. Immunostaining for a„|3 3 Integrin (VNR) 

1. Monoclonal antibody 23c6 (anti-a v |3 3 ). 

2. Fluorescently conjugated secondary antibody (e.g., Alexa Fluor 594 goat anti- 
mouse IgG; Molecular Probes, OR, USA). 

3. 0.5 |xg/mL of 4',6-diamidino-2-phenylindole (DAPI). 

2.7. Detection of F-Actin Rings 

1. 0.5% (v/v) Triton X-100 in PBS. 

2. Tetramethylrhodamine isothiocyanate (TRITC)-phalloidin (Sigma, Poole, UK). 

2.8. Resorption Pit Assay 

1. Ivory discs: Cut 200-jxm sections of 1-cm 2 blocks of elephant ivory using a 
Beuhler low-speed saw with wafering blade. The surface of the slices is pol- 
ished by rubbing vigorously with tissue paper, then the slices are punched 
into discs using a paper punch. The discs can be sterilized and stored in 95% 

2. 20% (w/v) Sodium hypochlorite solution. 

07/Rogers/89-100/F1 93 1 2/26/03, 10:45 AM 

I ■ I I I 


94 Coxon et al. 

3. Methods 

3. 1. Isolation and Culture of Rabbit Osteoclasts 

1. Euthanize 2- to 4-d-old rabbits under halothane. 

2. Remove all four limbs entirely, remove skin, and keep in PBS on ice. 

3. In a glass Petri dish, remove all tissue from the femorae, tibulae, ulnae, and radii 
with a disposable scalpel. Transfer the bones to PBS as soon as they are dissected. 

4. Create a mixed cell suspension in a glass Petri dish by mincing all isolated bones 
from one rabbit in a-MEM (approx 20 mL) using a scalpel. For the larger bones 
it is best to cut longitudinally first, then scrape out the marrow and the inside of the 
bones before mincing the remaining bone. It is important to perform this part of the 
procedure as quickly as possible, as the osteoclasts settle and adhere to the dish. 

5. Transfer the cell suspension and bone fragments to a 50-mL conical tube and 
vortex-mix vigorously for three 10-s bursts. Allow the bone fragments to settle 
for 1 min and then decant the cell suspension to a fresh tube. Add a-MEM and 
supplement with FCS to a final concentration of 10% (v/v) in a final volume of 
25 mL or 50 mL (see Table 1). This suspension should contain approx 1 x 10 8 
total cells. 

6. Plate out the mixed cell suspension into petri dishes (see Notes 1 and 2) or 
multiwell plates using the guidelines in Table 1. 

7. Incubate cultures overnight in 5% C0 2 at 37°C, then remove nonadherent cells by 
washing gently in PBS using a sterile, wide-bore pasteur pipet. The remaining 
adherent cells are mainly osteoclasts, prefusion mononuclear cells, and stromal cells. 

3.2. Purification of Rabbit Osteoclasts on Plastic 

1. After overnight incubation of the bone marrow suspension as described in Sub- 
heading 3.1., step 7, remove the medium containing nonadherent cells, then wash 
the plates gently with PBS using a sterile, plastic pasteur pipet. Three washes are 
usually sufficient to remove most of the nonosteoclastic cells (see Notes 3 and 4). 

2. If significant numbers of contaminating cells remain, incubate the remaining 
adherent cells for 5-10 min (or until the nonosteoclastic cells are released) in 
prewarmed 0.001% (w/v) pronase, 0.002% EDTA in PBS, at 37°C. 

3. Wash the plates four times in PBS and culture the remaining purified (typically 
>95% pure) osteoclasts and prefusion mononuclear cells in a-MEM supple- 
mented with 10% (v/v) FCS. The purified osteoclast cultures are typically about 
10-20% confluent, and each 10-cm Petri dish usually yields 100-200 jxg of cel- 
lular protein following lysis (see Note 5). 

3.3. Generating Large Numbers of Rabbit Osteoclast-Like 
Cells In Vitro 

1. After allowing mature osteoclasts to adhere to culture dishes overnight, as 
described in Subheading 3.1., remove the nonadherent bone marrow cells and 
pool with cells that have been removed by washing in PBS and pronase-EDTA 
digestion. Seed these cells into 10-cm diameter petri dishes (3 x 10 7 cells per 

07/Rogers/89-100/F1 94 I 2/26/03, 10:45 AM 

I ■ I I I 


10-cm Petri dish 

50 mL 

6-well plate 

25 mL 

24-well plate 

25 mL 

48-well plate 

25 mL 

96-well plate 

25 mL 

Rabbit Osteoclasts 95 

Table 1 

Recommended Density of Bone Marrow Cells for Preparing Cultures 

of Rabbit Osteoclasts (see Note 2) 

Volume per well (approx cell 
Culture vessel Total volume of cell suspension number) 

16 mL (3.2 x 10 7 ) 

2 mL (8 x 10 6 ) 

0.5 mL (2 x 10 6 ) 

0.3mL(1.2x 10 6 ) 

0.125 mL (0.5 x 10 6 ) 

dish) in a-MEM supplemented with 10% (v/v) FCS and containing 1 x 10 8 M 
l,25-(OH) 2 D 3 . 

2. Replace half of the medium [containing l,25(OH) 2 D 3 ] every 2 d. 

3. After 10 d, remove the stromal layer by washing extensively with PBS. This 
usually yields >95% pure multinucleated, TRAP-positive osteoclast-like cells 
(capable of resorbing bone mineral when the marrow cells are cultured on ivory 
discs); therefore further purification using pronase-EDTA is not usually required. 

3.4. Isolation of Rabbit Osteoclasts Using Immunomagnetic Beads 

1. Prepare a mixed cell suspension from the long bones of a neonatal rabbit as 
described in Subheading 3.1. 

2. Centrifuge the mixed cell suspension at 300g (10 min) and resuspend the cell 
pellet in 1.0 mL of undiluted 23C6 hybridoma supernatant for 30 min at 37°C. 

3. Centrifuge the cells at 300g (5 min), then wash in 0.1% (w/v) BSA in PBS and 
resuspend in 0.1% (w/v) BSA in PBS containing 2 x 10 7 magnetic Dynal beads 
conjugated to rat anti-mouse IgG. Incubate at 4°C on a rotating mixer for 20 min. 

4. Separate the VNR-positive from VNR-negative cells by placing in a Dynal mag- 
netic particle concentrator for 5 min. Wash the VNR-positive cells (osteoclasts) 
four times in 0.1% (w/v) BSA in PBS, separating the bead-coated cells for 1 min 
after each wash. Finally, place the VNR-negative fraction into the magnet to 
retrieve any lost beads, wash these beads, and add to the VNR-positive fraction. 

5. Count the number of purified osteoclasts using a hemocytometer. This technique 
typically yields approx 2 x 10 4 TRAP-positive multinucleated cells that are 
capable of resorption when cultured on a mineralized substrate (Fig. 3) (see Note 6). 

3.5. Characterization of Rabbit Osteoclasts 

3.5.1. Staining for TRAP 

1. Osteoclast cultures should be rinsed with PBS then fixed with 4% (v/v) formal- 
dehyde in PBS for 10 min prior to staining for TRAP. Fixed cells can then be 
stored at 4°C in PBS for up to 2 wk before staining for TRAP. 

07/Rogers/89-100/F1 95 1 2/26/03, 10:45 AM 

I ■ I I I 


96 Coxon et al. 

2. Prepare staining solution in glass vials by mixing solutions A and B described below 
(see Note 7). Once prepared, this staining solution should be used the same day. 

Solution A: 150 uL naphthol-AS-BI-phosphate stock, 750 jxL of veronal buffer, 
pH 10. 1 ; 0.45 mL of acetate buffer; 1 .35 mL of acetate buffer-100 mM tartrate. 
Solution B: 120 \xL of pararosanilin, 120 jxL of sodium nitrite (4% w/v). Filter 
staining solution through a 0.2-ixm filter before use. 

3. Incubate osteoclasts in filtered staining solution at 37 C C for 30-60 min. TRAP- 
positive cells metabolize the substrate to a red product that appears as granular 
staining throughout the cytoplasm of osteoclasts. Cells should be rinsed in PBS 
and can then stored in 70% (v/v) ethanol at 4°C for several weeks. 

4. Count the number of osteoclasts (TRAP-positive, multinucleated cells with >2 
nuclei) under bright field illumination at x20 magnification. Because mono- 
nuclear prefusion osteoclasts are also TRAP positive it is important to verify the 
number of nuclei (easily distinguished by negative contrast as unstained areas) 
within each cell when counting 

3.5.2. Fluorescence Immunostaining fora$ 3 Integrin (VNR) 

1. Rinse cells in PBS, then fix in 4% formaldehyde for 10 min. 

2. Incubate cells in 10% (v/v) FCS in PBS for 20 minutes. 

3. Incubate cells in undiluted 23C6 (anti-VNR) hybridoma supernatant for 30 min 
at room temperature. 

4. Wash three times in PBS containing 0.1% (v/v) FCS. 

5. Incubate with fluorescently labeled, secondary antibody at a dilution of 1:80 
(25 ng/mL) in PBS, for 30 min. 

6. To identify multinucleated cells, nuclei can be fluorescently stained by incubat- 
ing cells for 10 min with a 0.5 jig/mL solution of DAPI in PBS. 

7. Rinse cells with PBS then visualize using a fluorescence microscope equipped 
with a 20x or 40x objective and appropriate filters (Fig. IB). 

3.5.3. Detection of F-actin Rings 

1. Rinse cells in PBS then fix in 4% formaldehyde. 

2. Permeabilize cells with 0.5% (v/v) Triton X-100 in PBS for 20 min. 

3. Incubate with 0.5 [xg/mL of TRITC-phalloidin in PBS for 30 min. 

4. Rinse twice in PBS, then store in PBS at 4°C, protected from light. 

5. Visualise actin rings using a fluorescence microscope with appropriate filters 
(Fig. 4B) (see Note 8). 

3.5.4. Resorption Pit Assay 

1. Prepare osteoclasts as described in Subheading 3.1. or 3.3., and seed onto min- 
eral discs in 96-well plates. 

2. Allow seeded osteoclasts to adhere to mineral discs for 2 h then gently rinse the 
discs in PBS to remove the nonadherent cells (see Note 9). At this stage, add any 
agents further agents (e.g., drugs or cytokines) into fresh oc-MEM containing 
10% (v/v) FCS. 

07/Rogers/89-100/F1 96 1 2/26/03, 10:45 AM 

I ■ I I I 


Rabbit Osteoclasts 97 

3. At the end of the culture period (see Note 10), fix the cells on the discs in 4% (v/v) 

4. Immerse the discs in 20% (w/v) sodium hypochlorite solution, followed by vig- 
orous wiping with a tissue to remove cells. 

5. Visualize resorption pits using a reflected light microscope. Areas of resorption 
appear dark because the uneven surface of the disc scatters the light, whereas 
unresorbed, flat surfaces appear bright because they reflect light (Fig. 4C). Alter- 
natively, pits can be visualized using a conventional light microscope after stain- 
ing the discs with 0.5% (v/v) toluidine blue. 

6. The area of resorbed mineral can be quantitated using image analysis software. 
We use a Leitz Q500MC image analysis system (Leitz, Milton Keynes, UK) with 
Aphelion-based software developed in house (see Chapter 11, Subheading 3.5.). 
The cultures prepared as described above usually result in 0.5-1 mm 2 of resorbed 
mineral per disc. 

4. Notes 

1. The source of Petri dishes appears to influence the yield of osteoclasts. In our hands, 
tissue-culture grade Falcon 10-cm diameter Petri dishes produce the best results. 

2. It is difficult to obtain confluent monolayers of purified rabbit osteoclasts, as 
seeding the cells at higher densities than those indicated in Table 1 prevents 
attachment of the osteoclasts to tissue culture dishes. 

3. Extensive washing with PBS is often sufficient to remove contaminating stromal 
cells, which may lift off as a single layer. In these cases digestion with pronase- 
EDTA is unnecessary. However, cultures should be only gently rinsed with PBS 
using a wide-bore, plastic pasteur pipet, to avoid damaging the osteoclasts. 

4. Rabbit osteoclasts have a relatively long life span when cultured in vitro, even 
following purification. Although cell number gradually declines, some rabbit 
osteoclasts remain viable after more than 1 wk in culture, without the addition of 
exogenous growth factors or supplements other than FCS. 

5. Rabbit osteoclasts are extremely adherent to tissue culture plastic and difficult to 
remove enzymatically. Therefore, when preparing osteoclast lysates for Western 
blot analysis, we lyse the cells directly in the Petri dish. These lysates can be 
concentrated if necessary by centrifuging through a microconcentrator (we use 
12-kDa molecular mass cutoff). 

6. We typically use immunomagnetic separation to isolate osteoclasts for the prepa- 
ration of cell lysates (e.g., for Western blot analysis) following in vivo adminis- 
tration of pharmacologic agents. Although the VNR-positive osteoclasts with 
magnetic beads attached can be cultured on ivory discs, the resorptive function of 
these cells is typically 20% of that of osteoclasts that have not been separated 
using magnetic beads. 

7. A final tartrate concentration of 50 mM is used when staining for TRAP in rabbit 
osteoclast cultures, which is higher than that used for staining for TRAP in osteo- 
clasts from other species, for example, mouse (use a final tartrate concentration 

07/Rogers/89-100/F1 97 I 2/26/03, 10:45 AM 

I ■ I I I 


98 Coxon et al. 

8. F-actin-containing podosomes can also be observed in osteoclasts cultured on 
plastic or glass following staining with TRITC-phalloidin, but on these surfaces 
osteoclasts do not form a genuine "F-actin ring." 

9. When seeding osteoclasts onto ivory discs to assess resorptive activity, it is 
important to wash off the nonadherent cells gently after seeding, as subsequent 
resorption appears to be greatly reduced in the presence of the nonadherent cells. 
The nonadherent cells can be removed as little as 1 h after seeding cells onto 
ivory discs. This will result in a purer population, but lower yield, of osteoclasts. 

10. Rabbit osteoclasts adhere rapidly to ivory but, in our hands, do not begin to resorb 
until about 12 h after seeding. We routinely incubate cultures of rabbit osteo- 
clasts for 48 h to assess resorptive activity. Only approx 25-50% of the TRAcP- 
positive, multinucleated osteoclasts seeded onto ivory discs are active (i.e., 
exhibit actin rings) at any one time. 


1. Shakespeare, W., Yang, M., Bohacek, R., et al. (2000) Structure-based design of 
an osteoclast-selective, nonpeptide src homology 2 inhibitor with in vivo 
antiresorptive activity. Proc. Natl. Acad. Sci. USA 97, 9373-9378. 

2. Fisher, J. E., Rogers, M. J., Halasy, J. M., et al. (1999) Alendronate mechanism of 
action: geranylgeraniol, an intermediate in the mevalonate pathway, prevents 
inhibition of osteoclast formation, bone resorption and kinase activation in vitro. 
Proc. Natl. Acad. Sci. USA 96, 133-138. 

3. Coxon, F. P., Helfrich, M. H., van 't Hof, R. J., et al. (2000) Protein 
geranylgeranylation is required for osteoclast formation, function, and survival: 
inhibition by bisphosphonates and GGTI-298. /. Bone Miner. Res. 15, 1467-1476. 

4. Benford, H. L„ McGowan, N. W„ Helfrich, M. H., Nuttall, M. E., and Rogers, 
M.J. (2001) Visualization of bisphosphonate-induced caspase-3 activity in 
apoptotic osteoclasts in vitro. Bone 28, 465-473. 

5. Weidema, A. F., Dixon, S. J., and Sims, S. M. (2001) Activation of P2Y but not 
P2X(4) nucleotide receptors causes elevation of [Ca 2+ ] ; in mammalian osteoclasts. 
Am. J. Physiol. Cell Physiol. 280, C1531-C1539. 

6. Lees, R. L., Sabharwal, V. K., and Heersche, J. N. (2001) Resorptive state and 
cell size influence intracellular pH regulation in rabbit osteoclasts cultured on 
collagen-hydroxyapatite films. Bone 28, 187-194. 

7. Chikazu, D., Hakeda, Y., Ogata, N., et al. (2000) Fibroblast growth factor (FGF)- 
2 directly stimulates mature osteoclast function through activation of FGF recep- 
tor 1 and p42/p44 MAP kinase. /. Biol. Chem. 275, 31,444-31,450. 

8. Tezuka, K., Sato, T., Kamioka, H., et al. (1992) Identification of osteopontin in 
isolated rabbit osteoclasts. Biochem. Biophys. Res. Commun. 186, 911-917. 

9. David, J. P., Neff, L„ Chen, Y., Rincon, M„ Home, W. C, and Baron, R. (1998) 
A new method to isolate large numbers of rabbit osteoclasts and osteoclast-like 
cells: application to the characterization of serum response element binding pro- 
teins during osteoclast differentiation. /. Bone Miner. Res. 13, 1730-1738. 

07/Rogers/89-100/F1 98 1 2/26/03, 10:45 AM 


I ■ I I I 

Rabbit Osteoclasts 99 

10. Collin-Osdoby, P., Oursler, M. J., Webber, D., and Osdoby, P. (1991) Osteoclast- 
specific monoclonal antibodies coupled to magnetic beads provide a rapid and 
efficient method of purifying avian osteoclasts. /. Bone Miner. Res. 6, 1353-1365. 

11. Nesbitt, S., Nesbit, A., Helfrich, M., and Horton, M. (1993) Biochemical charac- 
terization of human osteoclast integrins. Osteoclasts express alpha v beta 3, alpha 
2 beta 1, and alpha v beta 1 integrins. /. Biol. Chem. 268, 16,737-16,745. 

12. Minkin, C. (1982) Bone acid phosphatase: tartrate-resistant acid phosphatase as a 
marker of osteoclast function. Calcif. Tissue Int. 34, 285-290. 

13. Van 't Hof, R. J., Tuinenburg-Bol, R. A., and Nijweide, P.J. (1995) Induction of 
osteoclast characteristics in cultured avian blood monocytes; modulation by 
osteoblasts and l,25-(OH)2 vitamin D3. Int. J. Exp. Pathol. 76, 205-214. 

07/Rogers/89-100/F1 99 1 2/26/03, 10:45 AM 

I ■ I I I 




2/26/03, 10:45 AM 

I ■ I I I 



Generating Human Osteoclasts 
from Peripheral Blood 

Afsie Sabokbarand Nicholas S. Athanasou 

1 . Introduction 

1. 1. Historical Perspective 

Osteoclasts are large multinucleated cells that are uniquely specialized for 
the function of lacunar bone resorption. For much of the previous century 
osteoclasts were thought to share a common progenitor cell with osteoblasts, 
bone-forming cells. Osteoclast formation occurring as a consequence of fusion 
of nonosteoblastic mononuclear precursor cells was suggested by a number of 
early investigators including Pommer (1883), Mallory (1912), La Coste (1923), 
and Hancock (1949) (for review see ref. 1). It was subsequently established by 
numerous studies that osteoclasts are derived from a hematopoietic marrow 
precursor (2). These studies included parabiosis experiments in which normal 
and affected (osteopetrotic/radiation-treated) littermates are linked by a com- 
mon circulation; these experiments established that the mononuclear osteo- 
clast precursor is present in peripheral blood (3,4). 

Following establishment of the hematopoietic origin of osteoclasts, there 
was considerable controversy as to the precise nature of the mononuclear 
osteoclast precursor. Neither mitotic nor amitotic division has been seen in 
osteoclasts, and osteoclast formation is thought to occur as a consequence of 
fusion of mononuclear precursors. As monocytes and macrophages fuse to form 
polykaryons that morphologically resemble osteoclasts, and osteoclasts are 
known to be phagocytic cells, it was suggested that osteoclasts form part of 
the mononuclear phagocyte system and that monocytes or at least mono- 
cyte-like cells could represent osteoclast precursors (5,6). There are numerous 
structural, functional, and immunophenotypic similarities between monocytes 
and osteoclasts. However, there are also significant differences, notably the 
inability of monocytes, macrophages, or their fused products, macrophage 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


08/Sabokbar/101-112/F1 101 1 2/26/03, 10:45 AM 

I ■ I I I 

102 Sabokbar and Athanasou 

polykaryons, to carry out lacunar resorption of a mineralized substrate, an 
essential functional defining characteristic of an osteoclast. On this basis, it 
was suggested that the osteoclast and its precursors could represent a unique 
type of phagocytic cell that was derived from a differentiation pathway that 
was distinct from that of monocytes and macrophages. 

Some of the cellular and humoral requirements for osteoclast forma- 
tion were identified in long-term hematopoietic (bone marrow or spleen) 
culture systems. In such cultures, it was found that osteoblasts or other 
specific bone-derived stromal cells are an absolute requirement for osteo- 
clast precursors of hematopoietic origin to differentiate into functional 
osteoclasts (7). In human bone marrow cultures it was noted that in the 
presence of 1,25-dihydroxyvitamin D 3 [l,25-(OH) 2 D 3 ], osteoclast-like 
multinucleated cells formed from colony-forming units for granulocytes 
and macrophages (8). It was subsequently shown that mouse monocytes 
and mononuclear phagocyte populations of extraskeletal tissue origin 
(alveolar macrophages) were found to be capable of osteoclast differen- 
tiation when these cells were cocultured with bone marrow stromal cells 
or osteoblast-like cell lines (9,10). As these rodent cells produce mac- 
rophage colony-stimulating factor (M-CSF) that does not react with the 
-®- human M-CSF receptor, it was found that in order to generate human- 

osteoclasts from monocytes and tissue macrophage populations, human 
M-CSF had to be added to cocultures of human monocytes and bone- 
derived stromal cells in addition to l,25-(OH) 2 D 3 (11). 

The necessity for contact between circulating or marrow-derived 
osteoclast precursors and bone stromal cells to generate osteoclasts in vitro 
suggested that osteoclast formation involved a ligand-receptor interaction 
between these cell types. This hypothesis was confirmed by the discovery 
of the receptor activator for nuclear factor kB ligand (RANKL), which is 
expressed by bone stromal cells, and RANK, which is expressed by osteo- 
clast precursors (including cells of the monocyte fraction) (12,13). This 
permitted in vitro methods of human osteoclast formation to be devised in 
which cultures of peripheral blood mononuclear cells (or tissue macroph- 
ages) alone, in the presence of RANKL and M-CSF, can be used for osteo- 
clast generation (14,15). 

Although M-CSF appears to be an absolute requirement for osteoclast 
formation, recent studies have shown that RANKL-induced osteoclast for- 
mation may not be the only means whereby bone-resorbing multinucleated 
cells are formed. It has been shown that cytokines (tumor necrosis factor- 
a[TNF-a] ± interleukin-a [IL-a]) can substitute for RANKL in generating 
osteoclasts from mouse marrow precursors under some circumstances (16). 
We have recently shown that this RANKL-independent mechanism is also 

08/Sabokbar/101-112/F1 102 I 2/26/03, 10:45 AM 

I ■ I I I 

Osteoclasts from Peripheral Blood 103 

capable of generating osteoclasts from human circulating precursors de- 
rived from the monocyte fraction (17,18). 

1.2. Defining Characteristics of Osteoclasts 

In employing culture systems of peripheral blood mononuclear cells (or 
other tissues) to generate osteoclasts, it is important to establish the specific 
defining characteristics of osteoclasts. A number of ultrastructural, cytochemi- 
cal, immunophenotypic, hormone receptor, and functional characteristics have 
been used to define a cell as an osteoclast (5). It should be noted that some of 
these markers are not exclusively expressed by osteoclasts. Cytochemical 
markers such as tartrate-resistant acid-fast phosphatase (TRAP), which is not 
expressed by isolated monocytes, can be expressed by mononuclear and multi- 
nucleated cells that form in cultures of monocytes and macrophages; these cells 
do not possess the ability to carry out lacunar resorption and express pheno- 
typic markers that are not found on osteoclasts, such as CD14, CD11/18, and 
HLA-DR. A useful osteoclast immunophenotypic marker is the vitronectin 
receptor CD5 1 (VNR) which is strongly (but not exclusively) expressed by 
osteoclasts (19). Calcitonin receptors (CTRs) and inhibition of osteoclast 
resorption by calcitonin are also thought to be specific for osteoclasts (20). 
-W~ Markers that have been used to identify osteoclasts are shown in Table 1. At 

the risk of stating the obvious, it should be noted that the only defining charac- 
teristic that permits a cell to be classified as an osteoclast is a functional one, 
that is, the ability of a cell to carry out lacunar resorption of a mineralized 
substrate. Monocytes, macrophages, and macrophage polykaryons are CTR 
negative and are not capable of lacunar resorption. 

2. Materials 

1. Ficoll-Hypaque (Pharmacia Biotech, UK). 

2. Phosphate-buffered saline (PBS). 

3. Acetic acid solution (5% [v/v] in H 2 0). 

4. Glass coverslips, 6-mm diameter. 

5. MACS CD 14 MicroBeads (Miltenyi Biotec). 

6. l,25-(OH) 2 D 3 (Solvay Dulpar, NL). 

7. Dexamethasone (Sigma-Aldrich Chemicals). 

8. Human M-CSF (R&D Systems Europe, Abingdon). 

9. 30 ng/mL of soluble RANKL (PeproTech, UK). 

10. TRAP kit (Sigma Aldrich Co., UK, Diagnostic Kit no. 386A). 

11. 1 mg/mL collagenase type I (Sigma-Aldrich Chemicals, cat. no. C-0130). 

12. 20 ng/mL of TNF-a (R&D Systems Europe, Abingdon). 

13. 70-jxm Cell strainer (Falcon, UK). 

14. Minimum essential medium (MEM) (Gibco). 

15. 10% Heat-inactivated fetal calf serum (FCS) (Gibco). 

08/Sabokbar/101-112/F1 103 1 2/26/03, 10:45 AM 

I ■ I I I 


Sabokbar and Athanasou 


Table 1 

Defining Criteria of Osteoclasts and Their Presumed Precursors 


Enzyme histochemistry 



Light microscopy: multinuclearity, location at 
resorbing sites 

Transmission electron microscopy": ruffled 
borders, clear zones 

Expression of TRAP fe , tartrate-resistant tri- 
nucleotide phosphatase,'' carbonic anhydrase 
isoenzyme type II,* cathepsin K ft . 

Expression of restricted range of leucocyte/ 
macrophage-associated antigens* (e.g., 
positive for CD13, CD15, CD45, CD68, CD51 
[VNR]; negative for CD11/18, CD14, 
HLA-DR, Fc and C3b receptors) 

Calcitonin receptor expression" 

Ability to form resorption lacunae on a bone 

Response to calcitonin/ability to bind calcitonin" 

F actin ring formation" 


''Osteoc last-associated but not specific. 

3. Methods 

3. 1. Isolation of Osteoclast Precursors from the Monocyte 
Population of PBMCs 

Human peripheral blood mononuclear cells (PBMCs) are isolated by Ficoll- 
Hypaque sedimentation and adherence from the peripheral blood of human 

1. Collect blood in EDTA-citrate tubes and dilute 1:1 with culture media. 

2. Layer over 5 mL of Ficoll-Hypaque and centrifuge at 510g for 20 min. 

3. Remove the mononuclear cell rich layer at the interface and wash twice in culture 

4. Resuspend the pellet in culture medium containing 10% serum. 

5. Lyse red cells using a 5% (v/v) acetic acid solution. 

6. Count cells in the resulting suspension using a hemocytometer. 

7. Add appropriate number of PBMCs (see either Subheading 3.2. or Subheading 
3.3.) to a 96-well plate containing either 4-mm diameter dentine slices or 6-mm 
diameter glass coverslips in culture medium with 10% serum. 

8. Ninety percent of the cells isolated from peripheral blood adhere to glass and are 
CD14-postitive monocytes (see Note 1). 



2/26/03, 10:45 AM 

I ■ I I I 

Osteoclasts from Peripheral Blood 



' orRANKI 
stromal cells 

Vj!.g. UMR 106 

Peripheral Blood Monocyte 

( 1)14' 





CD 14 

BR + 

Fig. 1. Schematic representation of the technique used for generating human osteo- 
clasts in vitro from peripheral blood mononuclear cells (PBMCs). M-CSF, macrophage 
colony stimulating factor; RANK, Receptor activator for nuclear factor-KB; RANKL, 
receptor activator for nuclear factor-KB ligand; sRANKL, soluble RANKL; TRAP, tar- 
trate-resistant acid phosphatase; VNR, vitronectin receptor; BR, bone resorption. 

3.2. Osteoclast Formation in Cocultures of Bone Stromal Cells 
and Monocytes (Fig. 1) 

1. Place dentine slices (4-mm diameter) or glass coverslips (6-mm diameter) in 96- 

well tissue culture plates. 

2. Add 2 x 10 4 UMR 106 cells (see Note 2) to each well and culture the cells on 

dentine slices and coverslips for 24 h in culture media containing serum. 

3. Add 1 x 10 5 PBMCs to each well and incubate for 2 h. 

4. Remove dentine slices and coverslips from the 6-mm wells, wash vigorously in 

culture medium to remove nonadherent cells, and place in a 24-well tissue cul- 
ture plate containing 1 mL of culture media supplemented with 10~ 7 M 1,25- 
(OH) 2 D 3 , 1(T 8 M dexamethasone, and 25 ng/mL of M-CSF. 

5. Maintain the cocultures for up to 21 d, during which time the medium (± added 

experimental factors, e.g., cytokines/growth factors) is replaced every 3 d. 

3.3. Generation of Human Osteoclasts in the Absence of Stromal 
Cells (Fig. 1) 

1. Add 5 x 10 5 PBMCs/well to wells of a 96-well tissue culture plate containing 
either dentine slices (4-mm diameter) or glass coverslips (6-mm diameter) as 
described in Subheading 3.2., step 1. 



2/26/03, 10:45 AM 

I ■ I I I 


106 Sabokbar and Athanasou 

2. After 2 h of incubation, remove the dentine slices and coverslips from the wells 
and wash vigorously in culture medium to remove nonadherent cells. 

3. Place slices or coverslips in 16-mm wells of a 24-well tissue culture plate con- 
taining 1 mL of culture medium containing serum supplemented with 30 ng/mL 
of soluble RANKL and 25 ng/mL of M-CSF ± 10~ 8 M dexamethasone (see Notes 
3 and 4). 

4. Incubate cultures for up to 21 d and replenish the culture medium containing the 
factors as in step 3 every 3 d. 

3.4. Assessment of Osteoclast Formation 

3.4.1. Cytochemical Assessment of Osteoclast Formation 

1. Carry out histochemical staining for TRAP, a marker of osteoclasts (28), using a 
commercially available TRAP kit (Sigma, Diagnostic Kit no. 386A) and counter- 
stain with hematoxylin. 

2. Alternatively stain cell preparations on coverslips immunohistochemically using 
an indirect immunoperoxidase method, with the monoclonal antibody 23C6 
(Serotec, Oxfordshire) to determine expression of VNR, an osteoclast-associated 

3.4.2. Functional Assessment of Osteoclast Formation 

Functional assessment of osteoclast formation is determined at the end of 
the culture period using a cortical bone or dentine slice resorption assay. 

1. Place dentine slices in 1 N NH 4 OH for 30 min and clean by ultrasonication to 
remove adherent cells. 

2. Wash slices with distilled water. 

3. Stain with 0.5% (v/v) toluidine blue for 3 min. 

4. Let slices air-dry. 

5. Examine by light microscopy for evidence of lacunar resorption. 

6. The extent of resorption is determined either by counting the number of discrete 
pits on each dentine slice or by measuring the percentage surface area resorbed 
on each dentine slice using an image analysis system (see Note 5). 

3.5. Preparation of Isolated Macrophages from Tissue Specimens 

It is possible to isolate mononuclear phagocyte osteoclast precursors from 
tissue specimens in much the same way as from suspensions of peripheral blood 
mononuclear cells. This is particularly useful in studying osteoclast formation in 
tissues affected by inflammatory and neoplastic disorders that are associated with 
bone resorption such as rheumatoid arthritis, aseptic loosening, and bone tumors. 

1 . Wash tissue specimens thoroughly with PBS before cutting into small fragments. 

2. Digest fragments in culture medium containing 1 mg/mL of collagenase type I 
for 30 min at 37°C. 

08/Sabokbar/101-112/F1 106 1 2/26/03, 10:45 AM 

I ■ I I I 


Osteoclasts from Peripheral Blood 107 

3. Follow this by a further 1-h incubation in 0.25% trypsin (Sigma- Aldrich Chemicals). 

4. Filter the digested tissue through a 70-fxm cell strainer and pellet cells in the 
filtrate by centrifugation at 2>00g for 10 min. 

5. Resuspend the cell pellet in culture medium containing 10% serum. 

6. Lyse red blood cells using a 5% (v/v) acetic acid solution and count the number 
of leukocytes in a hemocytometer. 

7. As there is a heterogeneous population of cells within each tissue specimen, 
add 1 x 10 5 CD14-postive-enriched cells (see Note 1) to 96-well plates containing 
either dentine slices or glass coverslips as described in Subheadings 3.1. and 3.2. 

3.6. Other Techniques of Osteoclast Generation from Peripheral 
Blood Mononuclear Cells, see Notes 6-8 

3.7. Advantages and Disadvantages of Using PBMCs 
for Generating Osteoclasts 

Generation of osteoclasts from PBMCs has distinct advantages. 

1. Osteoclast precursors can be obtained readily without the use of invasive tech- 

2. The technique permits the role of osteoclast formation in systemic osteolytic dis- 
orders such as Paget's disease and osteoporosis to be studied directly. 

3. It is possible to investigate the direct effect of cellular/humoral factors on an 
osteoclast or isolated population of osteoclast precursors. It is also possible to 
determine the role of other cell populations within peripheral blood mononuclear 
cell (e.g., T-cell subsets, granulocytes, etc.) on osteoclastogenesis (e.g., see the 
chapter by Flanagan and Massey, p. 113). 

There are, however, a few disadvantages of generating osteoclasts from 
peripheral blood monocytes and these include: 

1. The presence of relatively few mononuclear phagocyte osteoclast precur- 
sors cells in specimens of whole blood. Relatively large blood specimens 
are required to obtain adequate number of precursors for experiments (usu- 
ally about 40-50 mL of whole blood). 

2. There is individual and sex/age variability in the number of osteoclast 
precursors in peripheral blood; large numbers of specimens are required 
to carry out appropriate statistical analyses. 

3. In vitro generation of human osteoclasts is time consuming and labor 
intensive, as each experiment is carried out over a period of 3 wk dur- 
ing which the culture media (plus added factors) need to be replen- 
ished twice weekly. Please note that our experiences here differ from 
those of some other workers who obtain osteoclast formation within 2 
wk (see the chapter by Flanagan and Massey, this volume, p. 113). Each 
laboratory new to these techniques should try out the protocols given 
in this chapter and in the chapter by Flanagan and Massey and decide 
which culture period is best. 

08/Sabokbar/101-112/F1 107 I 2/26/03, 10:45 AM 

I ■ I I I 


108 Sabokbar and Athanasou 

4. Notes 

1 . To increase the number of these cells and to ensure that osteoclasts are generated 
from the CD14-postive fraction of circulating cells, isolated PBMCs can be incu- 
bated with CD 14 MicroBeads for 20 min at 4°C before being passed through a 
MACS magnetic cell separator (21) (see the chapter by Flanagan and Massey, 
p. 113, for full details). 

2. Other stromal cells that have been shown to support osteoclast formation from 
marrow or circulating cells include the cell culture lines ST2, SAOS-2, MG63, 
and cultured calvarial osteoblasts and human trabecular bone-derived cells 

3. Using the same methodology, Nicholson et al. (25) have recently shown that 
CD14-postive cells selected from human PBMCs can differentiate into active 
bone resorbing osteoclasts in vitro. They also noted that highly purified CD 14- 
positive cells do not express mRNA for RANKL or OPG but express mRNA for 
NF-kB, whereas PBMCs expressed mRNAs for all three. 

4. Although in this protocol RANKL is added throughout the culture, Nicholson et 
al. have recently described in detail the minimal exposure period for soluble 
RANKL required to generate osteoclasts from PBMC cultures (26). Contrary to 
findings of Lam et al. (27), their results suggested that early exposure of RANKL 
does not "prime" human PBMCs or enhance proliferation of RANKL responsive 
precursors. They have also shown that soluble RANKL is not required for each 
of the 3 wk of osteoclast formation, the second week of culture being the most 
crucial period of osteoclast generation from PBMCs. Moreover, they noted that 
incubation for as short as 1 h with soluble RANKL during the second week of a 
3-wk culture can produce bona fide bone resorbing osteoclasts (26). 

5. Resorption pits are observed as either individual small pits or large multilocular 
areas. As such, it is necessary to define a resorption pit as an excavation of the 
dentine surface with a clear rim of unchanged original surface between neighbor- 
ing excavations. Pit numbers are counted by a single blinded observer for all 
experiments. Measurement of the percentage area of resorption is a more accu- 
rate and reproducible method of quantifying the amount of lacunar resorption, 
particularly if this is extensive and there are numerous confluent areas of excava- 
tion on the dentine or bone slices. An image analysis method for measuring per- 
centage resorption is described in detail in the chapter by van 't Hof, p. 145). 

6. A RANKL-independent method of generating osteoclasts in vitro from human 
PBMCs has recently been described (17,18). In this system, 1 x 10 6 human 
PBMCs are cultured on dentine slices and coverslips in the presence of 25 ng/mL 
of human M-CSF for 3 d after which time they are transferred to new wells contain- 
ing M-CSF, 20 ng/mL of TNF-oc, and 10 ng/mL of IL-la for a further 17 d. This 
results in the formation of numerous TRAP-positive multinucleated cells (MNCs), 
some of which are capable of forming resorption pits that are generally smaller 
than those from PBMC cultures incubated with M-CSF and soluble RANKL. 

7. Based on the fact that osteoclasts (and other cells of the monocyte-macrophage 
lineage) are generated from a pluripotent haemopoietic progenitor that is known 

08/Sabokbar/101-112/F1 108 1 2/26/03, 10:45 AM 

I ■ I I I 


Osteoclasts from Peripheral Blood 109 

to express the antigen CD34, Matayoshi et al. (29) and Purton et al. (30) gener- 
ated human osteoclasts from long-term cultures of CD34-positive cells in periph- 
eral blood (this method is different from that described in the chapter by Flanagan 
and Massey, p. 113, where CD34-positive cells are directly isolated from blood 
marrow). As the percentage of CD34-positive cells in peripheral blood is very 
low (about 0.01-0.1%), each human donor was treated with granulocyte colony 
stimulating factor (G-CSF) for 3-5 d to mobilize the bone marrow CD34-posi- 
tive cell population into the peripheral circulation. At the end of each treatment 
period, peripheral blood leucocytes were harvested by single leukapheresis 
using a Spectra Cell Separator. The CD34-positive-enriched cell population 
was purified by magnetic activated cell sorting, using the MiniMACS separa- 
tion kit. CD34-positive cells (which did not express the stromal cell antigen, 
Stro-1) were seeded in 48-well culture plates and cultured in the presence of 
culture media containing GM-CSF, IL-1, IL-3, IL-6, and stem cell factor for 
6 wk. At the end of the culture period MNCs were purified by serum gradient 
fractionation and immunomagnetic selection with the anti-osteoclast antibody 
121F based on the methodology described in Chapter 6 by Collin-Osdoby et 
al., p. 65 (31). Characterization of the purified osteoclasts was carried out as 
described in Subheading 3.4. 
8. Faust et al. (32) have shown that when peripheral blood mononuclear cells are 
cultured at very high densities, these cells are capable of differentiating into bone 
resorbing osteoclasts in the absence of the addition of M-CSF or RANKL within 
2 wk. These cultures consist of adherent cells having low nuclearity (mainly 
mono-, di-, trinuclear, -10-20 \im in diameter). These preosteoclasts/osteoclasts 
were positive for TRAP, cathepsin K, VNR, and CTR and were capable of low 
level bone resorption. The authors suggest that cells express these osteoclast phe- 
notypic features form from PBMCs in this way owing to the initial high cell 
density (1.5 x 10 6 cells/cm 2 ; 100-fold higher than other PBMC culture systems) 
used in their assay (11,29). In this system it is likely that the essential differentia- 
tion factors required for osteoclast formation (such as M-CSF and RANKL) are 
supplied by other cell types present in peripheral blood, for example, monocytes 
and T and B lymphocytes. 


1. Hancox, N. M. (1972) The osteoclast, in The Biochemistry and Physiology of 
Bone, (Bourne, G. H., ed.) Academy Press New York, pp. 45-67. 

2. Marks, S. C, Jr and Popoff, S. N. (1988) Bone cell biology: the regulation of 
development, structure, and function in the skeleton. Am. J. Anat. 183, 1-44. 

3. Walker, D. G. (1973) Osteopetrosis cured by temporary parabiosis. Science 180, 875. 

4. Coccia, P. F., Krivit, W., Cervenka, J., et al. (1980) Successful bone-marrow trans- 
plantation for infantile malignant osteopetrosis. N. Engl. J. Med. 302, 701-708. 

5. Athanasou, N. A. (1996) Cellular biology of bone-resorbing cells. /. Bone Joint 
Surg. Am. 78, 1096-1112. 

08/Sabokbar/101-112/F1 109 1 2/26/03, 10:45 AM 

I ■ I I I 


110 Sabokbar and Athanasou 

6. Chambers, T.J. (1985) The pathobiology of the osteoclast. /. Clin. Pathol. 38, 

7. Takahashi, N., Yamana, H., Yoshiki, S., et al. (1988) Osteoclast-like cell forma- 
tion and its regulation by osteotropic hormones in mouse bone marrow cultures. 
Endocrinology 122, 1373-1382. 

8. Kurihara, N., Chenu, C, Miller, M., Civin, C, and Roodman, G. D. (1990) Iden- 
tification of committed mononuclear precursors for osteoclast-like cells formed 
in long term human marrow cultures. Endocrinology 126, 2733-2741. 

9. Udagawa, N., Takahashi, N., Akatsu, T., et al. (1990) Origin of osteoclasts: mature 
monocytes and macrophages are capable of differentiating into osteoclasts under 
a suitable microenvironment prepared by bone marrow-derived stromal cells. 
Proc. Natl. Acad. Sci. USA 87 7260-7264. 

10. Quinn, J. M., McGee, J. O., and Athanasou, N. A. (1994) Cellular and hormonal 
factors influencing monocyte differentiation to osteoclastic bone-resorbing cells. 
Endocrinology 134, 2416-2423. 

11. Fujikawa, Y., Quinn, J. M. W., Sabokbar, A., McGee, J. O'D., and Athanasou, N. 
A. (1996) The human mononuclear osteoclast precursor circulates in the mono- 
cyte fraction. Endocrinology 139, 4058-4060. 

12. Lacey, D.L., Timms, E., Tan, H-L., et al. (1998) Osteoprotegerin ligand is a 
cytokine that regulates osteoclast differentiation and activation. Cell 93, 165-176. 

13. Yasuda, H., Shima, N., Nakagawa, N., et al. (1998) Osteoclast differentiation fac- 
tor is a ligand for osteoprotegerin/osteoclastogenesis is identical to TRANCE/ 
RANKL. Proc. Natl. Acad. Sci. USA 95, 3597-3602. 

14. Quinn, J.M.W., Elliott, J., Gillespie, M.T., and Martin, T.J. (1998) A combination 
of osteoclast differentiation factor and macrophage-colony stimulating factor is 
sufficient for both human and mouse osteoclast formation in vitro. Endocrinology 
139, 4424-4427. 

15. Matsuzaki, K., Udagawa, N., Takahashi, N., et al. (1998). Osteoclast differentia- 
tion factor (ODF) induces osteoclast-like cell formation in human peripheral blood 
mononuclear cell cultures. Biochem. Biophys. Res. Comm. 246, 199-204. 

16. Kobayashi, K., Takahashi, N., Jimi, E., et al. (2000) Tumor necrosis factor alpha 
stimulates osteoclast differentiation by a mechanism independent of the ODF/ 
RANKL-RANK interaction. /. Exp. Med. 191, 275-286. 

17. Hirayama, T., Sabokbar, A., and Athanasou, N.A. (2000) Humoral factors act 
directly on circulating osteoclast precursors to control osteoclast formation. Calcif. 
Tissue Int. 66 (Suppl 1), 212. 

18. Kudo, O., Fujikawa, Y., Itonaga, I., Sabokbar, A., and Athanasou, N. A. (2002). 
Pro-inflammatory cytokine (TNFa/IL-la) induction of human osteoclast forma- 
tion. /. Pathol. 198, 220-227. 

19. Horton, M.A., Lewis, D., McNulty, K., Pringle, J.A.S., and Chambers, T.J. (1985) 
Monoclonal antibodies to osteoclastomas (giant cell bone tumours): definition of 
osteoclast specific cellular antigens. Cancer Res. 45, 5663-5669. 

20. Hattersley, G. and Chambers, T. J. (1989) Calcitonin receptors as markers for 
osteoclastic differentiation: correlation between generation of bone-resorptive 

08/Sabokbar/101-112/F1 110 1 2/26/03, 10:45 AM 

I ■ I I I 


Osteoclasts from Peripheral Blood 1 1 1 

cells and cells that express calcitonin receptors in mouse bone marrow cultures. 
Endocrinology 125 1606-1612. 

21. Massey, H.M. and Flanagan, A.M. (1999) Human osteoclasts derive from CD 14- 
positive monocytes. Br. J. Haematol. 106, 167-170. 

22. Blair, H.C., Sidonio, R.F., Friedberg, R.C., Khan, N.N., and Dong, S.S. (2000) 
Proteinase expression during differentiation of human osteoclasts in vitro./. Cell 
Biochem. 78, 627-637. 

23. Neale, S.D., Fujikawa, Y., Sabokbar, A., et al. (2000) Human bone-derived cells 
support formation of human osteoclasts from arthroplasty-derived cells in vitro. 
/. Bone Joint Surg. 82B, 892-900. 

24. Itoh, K., Udagawa, N., Matsuzaki, K., et al. (2000) Importance of membrane- or 
matrix-associated forms of M-CSF and RANKL/ODF in osteoclastogenesis sup- 
ported by SaOS-4/3 cells expressing recombinant PTH/PTHrP receptors. /. Bone 
Miner. Res. 15, 1766-1775. 

25. Nicholson, G.C., Malakellis, M., Collier, F.M., et al. (2000). Induction of osteo- 
clasts from CD14-positive human peripheral blood mononuclear cells by receptor 
activator of nuclear factor kappaB ligand (RANKL). Clin. Sci. 99, 133-140. 

26. Nicholson, G. C, Aitken, C. J., Hodge, J. M., et al. (2001). Limited RANKL 
exposure in vitro induces osteoclastogenesis in human PBMC. Bone 28 (Suppl), 

27. Lam, J., Takeshita, S., Barker, J. E., Kanagawa, O., Ross, P. F., and Teitelbaum, S. 
L. (2000) TNF-alpha induces osteoclastogenesis by direct stimulation of macroph- 
ages exposed to permissive levels of RANK ligand. /. Clin. Invest. 106, 1481-1488. 

28. Minkin, C. (1982) Bone acid phosphatase: tartrate-resistant acid phosphatase as a 
marker of osteoclast function. Calcif. Tissue Int. 34, 285-290. 

29. Matayoshi, A., Brown, C, DiPersio, J.F., et al. (1996) Human blood-mobilized 
hematopoietic precursors differentiate into osteoclasts in the absence of stromal 
cells. Proc. Natl. Acad. Sci. USA 93, 10,785-10,790. 

30. Purton, L.E., Lee, M.Y., and Torok-Storb, B. (1996) Normal human peripheral 
blood mononuclear cells mobilized with granulocyte colony stimulating factor 
have increased osteoclastogenic potential compared to non-mobilized blood. 
Blood 87, 1802-1808. 

31. Collin-Osdoby, P., Oursler, M.J., Webber, D., and Osdoby, P. (1991) Osteoclast- 
specific monoclonal antibodies coupled to magnetic beads provide a rapid and 
efficient method of purifying avian osteoclasts. /. Bone Miner. Res. 6, 1353-1365. 

32. Faust, J., Lacey, D.L., Hunt, P., et al. (1999) Osteoclast markers accumulate on 
cells developing from human peripheral blood mononuclear precursors. /. Cell 
Biochem. 72, 67-80. 

08/Sabokbar/101-112/F1 111 I 2/26/03, 10:45 AM 

I ■ I I I 



2/26/03, 10:45 AM 

Generating Human Osteoclasts In Vitro 
from Bone Marrow and Peripheral Blood 

Adrienne M. Flanagan and Helen M. Massey 

1. Introduction 

Osteoclasts derive from macrophage colony-stimulating factor (M-CSF)- 
dependent hemopoietic precursors that develop into cells that express the a v (3 3 
subunit of the vitronectin receptor (VNR) and the calcitonin receptor (CTR). 
The extracellular degradative process, known as bone resorption, is the hall- 
mark of the osteoclast, and includes removal of both the hydroxyapatite and 
organic components of the skeleton. For bone resorption to occur, osteoclasts 
form a subcellular space, referred to as an extracellular lysosome, into which 
they secrete acid and enzymes when they come into contact with either calci- 
fied bone or dentine but not with plastic or uncalcified collagen-based matri- 
ces. This subcellular space is dependent upon the formation of a "tight seal" by 
the osteoclast, a process involving rearrangement of the cytoskeleton into a 
characteristic F-actin ring structure (see the chapter by Nesbitt and Horton, this 
volume, for details). 

This chapter addresses the various means by which human osteoclasts can 
be generated in vitro and describes the conditions that we have found optimal 
for generation of osteoclasts from both human peripheral blood mononuclear 
cells and bone marrow hemopoietic precursors. These conditions are not always 
similar to those required for murine osteoclast formation, and readers are 
directed to the chapter by Takahashi et al., this volume, for this information. 

M-CSF-dependent precursors, capable of forming osteoclasts, exist in the 
bone marrow and peripheral and cord blood. From these hemopoietic sources, 
osteoclasts can be formed with relative ease and with good reproducibility pro- 
vided a few crucial conditions are adhered to. The phenotypes of osteoclasts in 
these assays are largely similar to those of osteoclasts formed in vivo (1 ,2), and 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


1 14 Flanagan and Massey 

in only one circumstance have we identified a phenotypic difference in osteo- 
clasts generated in vivo and in vitro (2), (see Table 1). Using hemopoietic 
precursors from different individuals, we have found that some variability is 
observed in the number of osteoclasts generated in vitro. However, we have 
not found that this prevents these osteoclast-forming in vitro assays from being 
used as a means of testing responses to various agents. For example, we have 
reproducibly found that: 17(3-estradiol suppresses osteoclast formation gener- 
ated from bone marrow cell cultures derived from post-menopausal women 
and men, but not in bone-marrow cultures generated from premenopausal 
women (3); prostaglandins are essential for osteoclasts generated from human 
bone marrow cell cultures (4); transforming growth factor (3 (TGF-(3) and 
interleukin (IL)-4 and IL-13 enhance the osteoclast-forming potential of 
peripheral blood hemopoietic precursors in a lymphocyte-rich microenviron- 
ment (5,6); and IL-4 and IL-13 consistently suppress osteoclast formation from 
CD14-derived peripheral blood mononuclear cells (PBMNCs) (6). Clearly, if 
near maximal osteoclast formation occurs in a particular culture in the pres- 
ence of M-CSF and the receptor activator for NFkB ligand (RANKL) (see 
Note 1) it is impossible to detect enhancement of osteoclast formation or bone 
resorption. This limitation of the human osteoclast forming in vitro assays is 
difficult to overcome because it is not possible to predict in advance which 
cultures give rise to small or large numbers of osteoclasts. 

2. Materials 

2. 1. For Cell Isolation 

1 . Ficoll-Paque (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, 

2. Phosphate-buffered saline (PBS). 

3. Various culture media are used successfully for the generation of human osteo- 
clasts including RPMI 1640, minimum essential medium (MEM) Eagle (Sigma- 
Aldrich Ltd., Gillingham, Dorset, UK), and MEM (Gibco Life Technologies, 
Paisley, Scotland, UK; see Note 2). The sera are supplemented with 10% batch- 
tested (see Note 3), heat-inactivated fetal or neonatal bovine serum (although 
15% of some sera are more effective at inducing osteoclast formation), 2 mM l- 
glutamine, 100 IU benzylpenicillin/mL, and 100 jxg of streptomycin/mL (this is 
referred to hereafter as complete medium). 

4. "Mister Frosty" vials (Nunc) or similar, for freezing cells. See Note 7. 

5. DNase (Roche Diagnostics Ltd., Lewes, E. Sussex, UK). See Note 28. 

2.2. For Osteoclast Cultures 

1. UMR 106.01 cells (a gift from Professor T. J. Martin, Melbourne, Australia) are 
grown in MEM. 

2. Devitalized bone slices or dentine discs (see Note 4). 

Human Osteoclasts from Marrow and Blood 1 15 

Table 1 

The Phenotype of the Osteoclast Precursor and Mature Resorbing 

form as They Develop from Peripheral Blood and Bone Marrow Precursors 

Blood Tissue 

Osteoclast precursor -» -» Nonresorbing osteoclast — » -» Resorbing osteoclast 

CTR - CTR + CTR + 

VNR - VNR + VNR + 

F-actin ring - F-actin ring + 

CDllb+ CDllb+ CDllb- 

CDllc+ CDllc+ CDllc+/- 

CD13+ CD13 + 

CD14+ CD14+ CD14- 

CD16- CD16- 

CD44 + CD44 + 

CD54 + CD54 + 

CD68 + CD68 + 

Peripheral blood mononuclear cells and bone marrow cells were plated on bone slices and 
cocultured with either UMR 106 cells in the presence of M-CSF or, in the abscence of UMR 106 
cells, with M-CSF and RANKL. The bone slices were removed from the cultures daily and 
assessed for the presence of cells simultaneously expressing the calcitonin receptor, using salmon 
125 I-calcitonin and other markers. In addition, cells forming F-actin ring structures were assessed 
for coexpression of other markers. The immunophenotype of the osteoclast, as determined by the 
expression of the calcitonin receptor or the F-actin ring, was similar in vivo to that observed in 
vitro for all markers except when osteoclasts were generated in vitro in the presence of soluble 
RANKL in the absence of a stromal cell population. Under these conditions CD lie was expressed 
by virtually all CTR-positive F-acting-positive resorbing osteoclasts in vitro whereas in the pres- 
ence of stroma CD1 lc is expressed only transiently by CTR-positive cells and its expression is 
lost when the F-actin ring is formed (see refs. 2 and 8). 

3. Cytokines/growth factors: TGF-(3, IL-4, and IL-13 (R&D Systems). M-CSF was 
a gift from the Genetics Institute (Boston, MA) and RANKL was a gift from 
Amgen (CA, USA) (see Note 5). Hydrocortisone, prostaglandin E 2 , and in- 
domethacin were purchased from Sigma Aldrich Ltd. (Poole, Dorset, UK). 

2.3. Immunomagnetic Cell Separation 

1. Fluorescein isothiocyanate (FITC)-labeled monoclonal antibodies to CD3 and 
CD 14, magnetic anti-FITC MicroBeads, CD34 Progenitor Cell Isolation Kit, 
positive selection columns, preseparation filters, and a magnet were all purchased 
from Miltenyi Biotec Ltd. (Bisley, Surrey, UK) FITC-labeled anti-CD19 was 
purchased from BD Biosciences (Oxfordshire, UK). 

2. Cell separating buffer was used throughout the magnetic cell sorting procedure 
and made as directed by the manufacturer (Miltenyi Biotec Ltd.). PBS, pH 7.2, 
was supplemented with 0.5% heat-inactivated fetal or neonatal bovine serum and 
2 mM ethylenediaminetetraacetic acid (EDTA; Sigma Aldrich Ltd.). 

1 16 Flanagan and Massey 

3. Methods 

3. 1. Preparation of Cells for Osteoclast Generation 

1. As sources of osteoclast precursors use: peripheral blood (obtained from volun- 
teers by venipuncture), cord blood (obtained with the consent of mothers from 
placentae postpartum, or bone marrow (aspirated from the posterior iliac crest of 
volunteers under either general or local anesthetic; see Note 6). Use heparin or 
EDTA to prevent coagulation. 

2. Dilute the blood and bone marrow 1:1 in PBS. Additional heparin or EDTA is not 

3. Separate mononuclear cells by centrifugation over Ficoll-Paque density gradi- 
ents at room temperature following the manufacturer's instructions. Gently layer 
35 mL of diluted peripheral blood or bone marrow over 15 mL of Ficoll-Paque 
and centrifuge at 1500 rpm for 30 min with the brake set "off." Keeping the same 
ratio, alter volume of blood and bone marrow to Ficoll-Paque as required. 

4. Remove the mononuclear cell layer from the Ficoll-Paque/plasma interface and 
wash in 10 mL of PBS at 1500 rpm for 10 min. The brake can be set at "on" on 
this occasion. Either use the cells immediately, or alternatively freeze them 
(approx 1 x 10 7 cells in 1 mL of serum containing 10% dimethyl sulfoxide 
[DMSO]). Freeze in a controlled manner by placing vials at -80°C overnight in a 
"Mister Frosty" filled with isopropanol, and store in liquid nitrogen for use at a 
later time (see Note 7). 

5. If the experiment is to be performed immediately, resuspend the pellet of cells in 
culture medium or cold cell-separating buffer for magnetic selection of mono- 
nuclear cell subpopulations as required. 

3.2. Generation of Osteoclasts from Human Peripheral Blood 
Mononuclear Cells Using UMR 106.01 Cells as a Source of RANKL 

This method is based on the original description of Fujikawa et al. (7). 

1. Plate 2 x 10 4 UMR 106.01 cells in 100 ^L of RPMI medium on each bone slice 
present in a 96-well plate. Incubate overnight at 37°C in a humidified atmosphere 
of 3% C0 2 /97% air (see Notes 8 and 9). 

2. Sediment 1 x 10 5 PBMNCs/well (prepared as described in Subheading 3.1.) in 
100 fxL of RPMI medium on the UMR cells. After 2-4 h remove the bone slices, 
one by one, from the wells using sterile forceps and wash the bone slices in PBS 
(containing antibiotics but not serum). Transfer each bone slice to a fresh well of 
a 96 well plate containing 100 jxl of RPMI medium/well (leaving behind the cells 
that have settled on the plastic). Add an additional 100 uL of medium/well con- 
taining 50 ng/mL of M-CSF (this gives a final concentration of 25 ng/mL). Incu- 
bate the cultures at 37°C in an incubator set to 3% C0 2 /97% air. (see Note 8 and 9). 

3. Feed the cultures twice per week (see Note 11) by removing half of the medium 
and adding 100 jxL of fresh medium containing M-CSF (25 ng/mL). Terminate 
the cultures as required; osteoclasts should appear in small numbers by d 3 and 
large areas of bone resorption should be present by d 10-14 that are rarely sig- 
nificantly increased after d 14. 

Human Osteoclasts from Marrow and Blood 1 1 7 

3.3. Generation of Human Osteoclasts from Washed PBMNCs 
in the Presence of M-CSF and RANKL (2,5,6) (see Note 9) 

1. Prepare PBMNCs as described in Subheading 3.1. 

2. Place bone slices in wells of a 96-well plate. 

3. Plate 2 x 10 5 PBMNCs in 100 ,uL of MEM, without any added cytokines, on the 
bone slices. Incubate the cells at 37°C at 5%C0 2 -95% air (see Note 10) for 2-4 h. 

4. Remove the plate from the incubator. Remove the bone slices, one by one, from 
the wells using sterile forceps and wash the bone slices in PBS (containing anti- 
biotics but not serum). 

5. Place the washed bone slices in wells of a new 96-well plate containing 100 jxL 
of MEM. 

6. Add an additional 100 [iL of medium (makes a total volume of 200 u,L) contain- 
ing 50 ng/mL of M-CSF (see Note 11) and 60 ng/mL of RANKL to each well. 

7. Place the tissue culture plate in the incubator (d 1) at 37°C in a humidified atmo- 
sphere of 5% C0 2 -95% air and remove for feeding on d 4 or 5 (feed 1; see Note 
12). Feeding involves removing 100 u,L of medium from each well and replacing 
this with 100 u,L of fresh medium containing 25 ng/mL of M-CSF and 30 ng/mL 
of RANKL. 

8. Feed the cultures again 3 or 4 d later (feed 2: d 7 or 8) as described in step 7. 

9. Feed again 3 or 4 d later (feed 3: d 10 or 1 1) as described in step 7. 
10. Stop the experiment on d 14 or as required (see Notes 13 and 14). 

3.4. Generation of Human Osteoclasts from CD14-Positive Cells 
Selected from PBMNCs (8) 

1. Prepare PBMNCs as in Subheading 3.1. 

2. Select CD14-positive cells by adding 10 \iL of anti-CD 14-FITC to 1 x 10' 
PBMNCs resuspended in 100 uL of cell separating buffer. Cell separating buffer 
is used throughout the magnetic cell sorting procedure and is made as directed by 
the manufacturer (Miltenyi Biotec Ltd.). In brief, PBS, pH 7.2, is supplemented 
with 0.5% heat-inactivated fetal or neonatal bovine serum and 2 mM EDTA 
(Sigma Aldrich Ltd.). Mix the cells and antibody well using a pipet and then 
incubate the mixture in the dark for 10 min at 6°C. MiniMACS columns can 
separate 10 3 — 10 7 labeled cells from a total population of 2 x 10 8 cells but larger 
scale selections are possible with different columns. 

3. Wash the labeled PBMNCs in cold cell- separating buffer and centrifuge at 1500 rpm 
for 10 min. Resuspend the pellet in the 10 p,L of separating buffer. Add 10 |xL of anti- 
FITC MicroBeads per 1 x 10 7 PBMNCs and incubate static for 15 minutes at 6°C 
in the dark. 

4. Repeat the washing step, resuspend the cell pellet in 500-1000 u,L of cell sepa- 
rating buffer, and pass the PBMNCs through a primed magnetic cell separation 
column. The CD14-positive cell fraction is held within the column, while the 
CD14-negative cells pass through. CD14-positive cells are then eluted from the 

1 18 Flanagan and Massey 

column on its removal from the magnetic field. The selected cells are expected to 
be at least 97% pure, and this can be assessed using a flow cytometer (see Note 15). 

5. Place bone slices in wells of a 96-well plate. 

6. Plate 5 x 10 4 -1 x 10 5 CD14-positive PBMNCs in 100 \iL of MEM medium on 
the bone slices. 

7. Continue with the culture as described in Subheading 3.3., steps 6-10. 

3.5 Osteoclast Formation from Unwashed (Lymphocyte-Rich) Human 
PBMNC Cultures in the Presence of Either I L-4 orlL-13(6) to In- 
crease Osteoclast Formation (see Notes 16 and 20) 

1. Isolate PBMNCs as described in Subheading 3.1. 

2. Place bone slices in wells of a 96-well plate. 

3. Plate 2 x 10 5 PBMNCs in 100 uL of MEM on the bone slices. 

4. Add 100 u,L of MEM containing 50 ng/mL of M-CSF and 2 ng/mL of IL-4 or 0.2 
ng/mL of IL-13 to each well (see Notes 17 and 18). 

5. Place the tissue culture plate in the incubator at 37°C in a humidified atmosphere of 
5% C0 2 -95% air (d 1) and remove for feeding on d 4 or 5 (feed 1 ; see Note 12). This 
involves removing 180 uL of medium and replacing this volume with fresh medium 
(see Note 19). This fresh medium contains 25 ng/mL of M-CSF and 30 ng/mL of 
RANKL. IL-4 or IL-13 is not added at this time; they are added only on d 1. 

6. Feed the cultures again 3 or 4 d later (feed 2: d 7or 8) but this time remove only 
100 u,L of the medium and replace with 100 uL of fresh medium containing 
25 ng/mL of M-CSF and 30 ng/mL of RANKL. 

7. Feed again 3 or 4 d later (feed 3: d 10 or 11). 

3.6. Osteoclast Formation from Unwashed (Lymphocyte-Rich) 
Human PBMNC Cultures in the Presence of TGF-fi (5) (see Notes 20 
and 21) 

1. Isolate PBMNCs as described in Subheading 3.1. Place bone slices in wells of a 
96-well plate. 

2. Plate 2 x 10 5 PBMNCs in 100 \iL of MEM medium on the bone slices. 

3. Add 100 uL of MEM medium containing 50 ng/mL of M-CSF and 20 ng/mL of 
TGF-|3 to each well to give a final concentration of 25 ng/mL of M-CSF and 
10 ng/mL of TGF-|3. Do not add RANKL at this point (see Notes 16 and 21). 

4. Place the tissue culture plate in an incubator at 37°C in a humidified atmosphere 
of 5% C0 2 -95% air (d 1) and remove for feeding on d 4 or 5 (feed 1; see Note 
12). This involves removing 100 u.L of medium and replacing with 100 uL of 
fresh medium containing 25 ng/mL of M-CSF and 60 ng/mL of RANKL. Do not 
add TGF-(3 at this time; it is added only on d 1; (see Note 21). 

5. Feed the cultures again 3 or 4 d later (feed 2: d 7 or 8). Remove 100 u.L of the 
medium and replace with 100 u.L of fresh medium containing 25 ng/mL of M- 
CSF and 30 ng/mL of RANKL. 

6. Feed again 3 or 4 d later (feed 3: d 10 or 1 1) as described in step 5. 

7. Stop the experiment on d 14 or as required (see Notes 13 and 14). 

Human Osteoclasts from Marrow and Blood 1 19 

3.7. Generation of Osteoclasts from Human Bone Marrow, 
General Considerations 

Generating osteoclasts from human bone marrow is extremely efficient, as 
marrow contains high numbers of early hemopoietic precursors. However, the 
effort of obtaining bone marrow on a regular basis in large volumes prohibits 
most research laboratories from using it frequently for the generation of human 

Three protocols are given below: one in which a stromal osteoblastic/fibro- 
blastic supporting cell layer is used as source of RANKL and in which no 
endogenous RANKL is added (two-phase culture system), a second in which 
osteoclasts are generated from nonadherent bone marrow M-CSF-dependent 
cells in the presence of recombinant human or murine RANKL, and a third in 
which purified early hemopoietic precursors (CD34-positive cells) are used to 
generate human osteoclasts in the presence of recombinant RANKL, M-CSF, 
and other cytokines. 

3.8 Generating Human Osteoclasts from Whole Human Bone 
Marrow Without Addition of RANKL 

This methodology was developed prior to the cloning of RANKL (13). It 
had been anticipated for some time that an osteoclast differentiation factor existed 
and that it was membrane bound (14). This method is reproducible but the draw- 
back is that osteoclast formation takes 3-4 wk. This was the first in vitro model 
in which 17(3-estradiol was found to suppress osteoclast formation (3,15). 

1 . Bone marrow is aspirated from the posterior iliac crest of volunteers under either 
local or general anesthetic (see Note 6). 

2. Prepare bone marrow mononuclear cells over a Ficoll-Paque density gradient as 
described in Subheading 3.1. 

3. Plate the cells in tissue culture flasks (10 5 cells/cm 2 of tissue culture flask) either 
in MEM or RPMI supplemented with M-CSF (final concentration 5 ng/mL) or in 
the absence of M-CSF (see Note 22). 

4. Feed the cultures on d 5 and d 10 by removing half of the medium and replacing 
this with the same volume of fresh medium containing 5 ng/mL of M-CSF and 
10 -6 M of hydrocortisone. Replace the non-adherent cells removed when feeding; 
this is achieved by centrifuging the spent medium and resuspending the cellular 
pellet in the fresh medium being added to the flasks. 10~ 6 M of hydrocortisone is 
present throughout the whole culture period (phases I and II; see Note 23). 

5. Maintain the culture until the cells become confluent; this occurs between d 9-14 
(see Note 24). 

6. Harvest the cells in the flasks using trypsin-EDTA. Centrifuge the cells to give a 
pellet and resuspend the cells in fresh medium and plate on bone slices (high 
density) at 10 5 cells/well in a 96- well plate. This is phase II of the culture. 

7. Add M-CSF (final concentration 50 ng/mL) to the cultures (see Note 22). 

120 Flanagan and Massey 

8. Terminate the cultures as described in Notes 13 and 14 for the PBMNC cultures. 
Osteoclasts and VNR- and CTR-positive cells appear on d 5-7, F-actin ring struc- 
tures appear 48 h later, and bone resorption commences almost immediately 
thereafter (see Note 25). 

3.9 Generation of Human Osteoclasts from M-CSF-Dependent 
NonAdherent Hemopoietic Precursors (see Note 26) 

1. To obtain a M-CSF-dependent population, prepare the mononuclear cells over 
Ficoll-Paque as described in Subheading 3.1. 

2. Remove the lymphocytes by magnetic selection using CD3 and CD 19 antibod- 
ies, following the protocol described in Subheading 3.4. Test for remaining lym- 
phocyte contamination using a flow cytometer (see Note 15). 

3. Plate the bone marrow mononuclear cells overnight in medium in a tissue culture 
flask (75 cm 2 ). The cell density is not critical at this point but approx 10 7 cells are 
plated in a 75-cm 2 tissue flask 

4. The next day harvest the nonadherent cells by centrifugation and resuspend the 
pellet in either complete MEM or RPMI medium. Discard the adherent cells. 

5. Plate the nonadherent cells on bone slices (10 5 cells/well of a 96- well plate; see 
Note 27) in the presence of M-CSF and RANKL (final concentrations 25 ng/mL 
and 30 ng/mL, respectively). Do not add hydrocortisone. 

6. Feed the cultures every 3-4 d (see Note 12) by removing half of the medium and 
replacing it with fresh medium and cytokines. 

7. To ensure that the hemopoietic precursors are truly M-CSF-dependent (and not 
surviving because stromal cells are present), control cultures in the absence of 
M-CSF should also be included. After 10 d all of the cells in these control cul- 
tures should be dead. 

8. Terminate the experiment at the time required (see Note 13); osteoclasts should 
be present on d 10 and extensive bone resorption by d 14 (see Note 14). 

3. 10 Generation of Osteoclasts from CD34-Positive Cells 
Selected from Bone Marrow 

1 . Aspirate bone marrow from volunteers into a heparinized container. 

2. Dilute 1 : 1 with serum-free RPMI containing 100 U/mL of DNase and shake gen- 
tly for 45 min at room temperature (see Note 28). 

3. Separate the mononuclear cells from the red blood cells using Ficoll-Paque, as 
described in Subheading 3.1. 

4. Select CD34-positive cells using the CD34 Progenitor Cell Isolation Kit, using 
the manufacturer's instructions (steps 5-10 below). 

5. Add 100 u.L of FcR Blocking Reagent per 1 x 10 8 total cells in 300 [xL of cell 
separating buffer C. 

6. Label cells by adding 100 uL of CD34 MicroBeads per 1 x 10 8 total cells. Mix 
well and incubate for 30 min at 6°C. 

7. Wash the cells in cell separating buffer at 1500 rpm for 10 min and resuspend the 
cell pellet in 500-1000 uL of buffer. 

Human Osteoclasts from Marrow and Blood 121 

8. Pass the mononuclear cells through a magnetic cell separation column via a 30-jJ.m 
filter to remove any clumps. This filtering process is used only when separating 
CD34-positive cells from bone marrow. The CD34-positive fraction is held 
within the column, while the CD34-negative cells pass through. 

9. Elute the CD34-positive cells from the column by removing it from the magnetic field. 

10. The purity of the selected cells can be tested by incubating 1 x 10 6 selected cells 
with an anti-CD34 monoclonal antibody labeled with phycoerythrin (PE) 
(Miltenyi Biotec Ltd.) and analyzed using a flow cytometer (see Note 15). 

11. Place bone slices in wells of a 96-well plate. 

12. Plate 2 x 10 4 CD34-positive mononuclear cells in 100 jxL of complete MEM or 
RPMI medium on the bone slices. 

13. Add 100 ixL of MEM-RPMI medium containing 50 ng/mL of M-CSF and 60 ng/mL 
of RANKL to each well (see Note 29). 

14. Proceed as described in Subheading 3.3., steps 7-10. 

4. Notes 

1. We define maximal osteoclast formation as resorption of 80-100% of the bone 
slices (3x3 mm). This equates to approx 600 VNR-positive cells which gener- 
ally occurs on d 10-14 of culture. 

2. The presence or absence of phenol red appears to make no significant difference 
to the outcome of peripheral blood or bone marrow experiments. 

3. The serum should be batch-tested (companies will hold sera batches for you for 
several weeks). A large order for the serum of your choice should then be placed. 
In this way, it is necessary only to batch-test sera approx every 2-3 yr. 

4. Devitalized bovine bone and dentine can be used as a substrate for human osteo- 
clast resorption. The preparation details of dentine slices are described in Chap- 
ter 11 by van 't Hof, this volume. Bone slices are prepared from bovine cortical 
bone; the bone is purchased fresh from the local butcher. It should be frozen at 
-20°C if not prepared immediately. Preparing the bone involves removing all of 
the adherent soft tissue with a sharp knife. The bone is cut transversely on a band 
saw and the fatty marrow is then removed. Subsequently, the rings of bone are 
cut into three or four segments that can be placed on the diamond blade bone saw 
from which wafers (1 mm thick) can be sectioned. Square bone slices (3x3 
mm 2 ) can be cut from the wafers using a scalpel or sharp, fine scissors. Dentine 
can be processed in the same way. Bone or dentine slices can be sterilized by UV 
light (in tissue culture hood cabinet, minimum of 30 min), or by storage in 75% 
ethanol. Rehydrate slices in PBS before use. The advantage of using square slices 
of substrate means that it is possible to visualize the cells on the plastic surround- 
ing the bone or dentine during the experiment; this cannot be done if discs of the 
substrate are employed as they fill the well. Plating cells in some additional wells 
without substrate overcomes this problem. 

5. M-CSF is synthesized in membrane-associated and secreted forms and can either 
be added to the osteoclast-forming cultures as a recombinant soluble molecule or 
can be present endogenously in the cultures (16). M-CSF is essential for human 

122 Flanagan and Massey 

osteoclast formation, but murine M-CSF does not mediate an effect on human 
cells (7,13). In contrast, human M-CSF induces osteoclast formation from murine 
hemopoietic precursors. There have been reports that vascular endothelial growth 
factor (VEGF) can substitute for M-CSF. However, we have been unable to 
reproduce this work. In our studies, we have used human M-CSF gifted from 
Genetics Institute (Boston, MA, USA), but human recombinant M-CSF is also 
commercially available from several companies, in particular R&D Systems 
and Peprotech (both in the UK). 

RANKL is a membrane-bound molecule, now available as a recombinant soluble 
molecule (for review see ref. 14). It induces osteoclast differentiation from M- 
CSF-dependent precursors. It also enhances osteoclast activity, thereby increas- 
ing bone resorption (17). However, this function is difficult to assess in 
recruitment assays because osteoclast formation continues throughout most of 
the experiment. RANKL is not species specific. In our work, we have used 
recombinant human RANKL, obtained as a gift from Amgen: however, RANKL 
is also commercially available from several companies, in particular Peprotech 
and Insight Biotechnology. It is advisable to batch-test and titrate different lots 
for biological activity. If large numbers of experiments are to be carried out, it is 
recommended that a large quantity is ordered, as this allows for considerable cost 
savings to be achieved. 

6. Local ethical committee approval should be obtained before human blood or bone 
marrow is taken and used in experiments. Informed consent should be obtained 
from all individual donors and from mothers before obtaining cord blood. 
Approximately 10 7 — 2 x 10 7 PBMNCs is generally obtained following Ficoll den- 
sity separation of 10 mL of peripheral blood. The number of mononuclear cells 
obtained from a bone marrow aspirate is considerably more variable and ranges 
between 10 7 and 10 8 mononuclear cells from a 10-mL aspirate. To obtain large 
numbers of PBMNCs (>10 8 ) it may be possible to obtain "blood packs," that is, 
the buffy coat from a single blood donation from an individual. Most blood trans- 
fusions are now given with leukocyte-depleted blood and many blood transfu- 
sion services will make excess or outdated blood packs available for research. 
Cells should be prepared over Ficoll-Paque as described in Subheading 3.1. This 
large volume of cells is useful for titration studies and experiments in which 
many different factors are tested. It is important to note that all cells are derived 
from an individual donor 

7. Up to 20% of the frozen sample of bone marrow and peripheral blood does not 
survive the freezing process. Apart from this initial loss, however, osteoclasts are 
formed in the normal fashion. To maximize cell viability it is important to thaw fro- 
zen cell samples quickly in warm water and immediately wash in culture medium. 

8. In the original description of the method of generating osteoclasts from periph- 
eral blood (7), bone slices on which cells were plated were cultured in wells of a 
24-well plate in an incubator set at pC0 2 of 5%. This procedure permitted the 
UMR cells to continue to grow and provided ample medium for these highly 
metabolic cells. We suspect that as a result of having a large volume of medium 

Human Osteoclasts from Marrow and Blood 123 

in the wells of 24-well plates, the pH of the medium remains sufficiently high 
(alkaline) to permit osteoclast formation. However, we made minor modifica- 
tions to this initial protocol. We found that if the bone slices bearing the plated 
mononuclear cells were transferred from a 96 well plated into fresh wells of a 96- 
well plate, instead of a 24-well plate, and the cultures were maintained in an 
incubator set at apC0 2 of 3%, osteoclasts formed efficiently (8). The reason for 
doing this was to prevent acidification of the medium, as it is well documented 
that osteoclast formation requires alkaline conditions (see the chapter by Hoebertz 
and Arnett, this volume). Ninety-six-well plates were chosen over 24-well plates, 
as fewer reagents are used and this reduces the cost of the experiments, and sec- 
ond, we find that microorganism contamination of the cultures occurs less fre- 
quently in 96- well plates. Our study and manipulation of the pC0 2 underscores 
the importance of pH in osteoclast formation. See also Note 10. 
9. Osteoclasts can be generated by culturing unfractionated PBMNCs in the pres- 
ence of M-CSF and RANKL. However, removal of the nonadherent population 
(mainly lymphocytes) by washing the bone slices 2-4 h (even up to 24 h) after 
the cells have been plated results in enhanced osteoclast formation (5,6). This 
washing procedure is tedious and is associated with increased risk of microor- 
ganism contamination. We have found that the addition of IL-4 or TGF-|3 to lym- 
phocyte-rich PBMNC osteoclast-forming cultures eliminates the need to remove 
nonadherent PBMNCs (see Subheadings 3.5. and 3.6.) (5,6). However, the par- 
ticular procedure that is chosen for generating human osteoclasts will depend on 
the question being investigated. 

10. The pH of the culture medium is critical for both osteoclast formation and bone 

resorption. This point cannot be overemphasized: Osteoclasts will not form or, if 
they form, they will not resorb, irrespective of the starting population of M-CSF- 
dependent cells and the concentrations of M-CSF and RANKL if the pH of the 
medium is not optimized. Arnett and co-workers have reported on this in detail 
using murine cultures (this volume). To date, no such detailed analysis of pH has 
been reported using human cultures. However, consistent with the fact that pH is 
determined by cell density (the higher the cell density, the more acidic the me- 
dium becomes) and the concentration of C0 2 , we have found that these two param- 
eters are crucial for the success of the human osteoclast-generating in vitro models 
described in this chapter. Also, ensure that there is sufficient water in the incubator 
at all times as failure to do so will alter the pC0 2 - Unlike the procedure described 
by Hoebertz and Arnett, this volume, we do not acidify our human osteoclast cul- 
tures by adding HC1, because we find that the pH of the medium becomes reduced 
sufficiently during the culture period to activate the osteoclasts to resorb. Regular 
monitoring of the concentration of the pC0 2 (twice per week) in the culture incuba- 
tors is performed using a Bacharach Fyrite (Leec Ltd., Nottingham, 
Nottinghamshire, UK). In all of our cultures, unless otherwise stated (Subheading 
3.2., Note 8), the C0 2 concentration in the incubator is 5%. 

1 1 . The final concentration of M-CSF in human osteoclast forming experiments is 
25 ng/mL, although this may depend on the source of the product. Titration 

124 Flanagan and Massey 

experiments may need to be performed to assess biological activity when a new 
batch or supplier is used (see also Note 5). When 100 jxL of medium containing 
M-CSF is first added to the cultures the concentration of M-CSF is 50 ng/mL 
because it is diluted in 100 jxL of medium already present in the culture wells. On 
feeding, 100 [iL of medium is removed from the wells of a 96 well plate and 100 \xL 
is replaced; this 100 fxL of fresh medium contains 25 ng/mL of M-CSF (not 50 
ng/mL) because it is considered that the cytokine in the medium that has not been 
removed will not have degraded significantly. This same presumption is made 
for other cytokines. 

12. The medium in osteoclast forming cultures is replaced every 3-4 d. As a rule, if 
an experiment is set up on Monday or Thursday it is fed on Thursdays and Mon- 
days. If an experiment is set up on Tuesday or Friday it is fed on Fridays and 
Tuesdays. If an experiment is set up on Wednesday, it is first fed on Monday and 
subsequently on Thursdays and Mondays. 

13. Experiments are stopped in different ways depending on which information is 
required. For assessment of osteoclast numbers using tartrate resistant acid phos- 
phatase (TRAP) or the VNR as markers and for assessment of bone resorption, 
bone slices are removed from the culture medium, washed in PBS to remove cell 
debris, and then left to air-dry. Fixation can be deferred until the immunohis- 
tochemistry or enzyme histochemistry is to be performed. The cells are fixed in 
10% formalin for TRAP staining and in cold acetone for labeling with antibody 
23c6 (for VNR). Bone and dentine substrates and plastic discs can be stored fro- 
zen following drying if histochemistry cannot be performed immediately. Bone 
resorption can be visualized under reflected light or by scanning electron micros- 
copy by either staining with toluidine blue or sputter-coating with gold, respec- 
tively. Fixation of cells for confocal microscopy requires that the cells are placed 
in fixative immediately. However, the type of fixation varies depending on the 
antibody being used and may require optimization for new untested antibodies. 
Details of these techniques are described in Chapter 1 1 by van 't Hof and Chapter 
19 by Nesbitt and Horton, this volume. 

14. In our experience, if large numbers of osteoclasts (approx 200 VNR-positive cells 
per bone slice) have not formed by d 10 and bone resorption has not occurred by 
d 14, the chance of successful generation of human osteoclasts is extremely low. 
However, occasionally osteoclasts can develop up to d 21 but this indicates that 
the conditions for osteoclast formation are suboptimal and it is therefore worth 
spending time optimizing the culture conditions. 

15. This is achieved by selecting fluorescently labeled cells using a fluorochrome 
conjugated primary antibody, for example, HPCA-2 (Becton Dickinson Ltd.). 
The analysis must include negative controls, including unselected PBMNCs and/ 
or unselected PBMNCs previously incubated with a FITC-conjugated, isotype- 
matched control antibody (Dako Ltd., Ely, Cambridgeshire, UK). 

16. The role of the immune system is intricately involved in the regulation of osteo- 
clast formation (9,10). We were interested in identifying which lymphocyte-pro- 
duced molecule(s) might suppress osteoclast formation and the mechanism by 

Human Osteoclasts from Marrow and Blood 125 

which this occurs and found that macrophage-deactivating molecules (TGF-|3, 
indomethacin [unpublished data], IL-4, and IL-13) block the suppressive effect 
that lymphocytes exert on human osteoclast formation from PBMNCs (5,6). The 
addition of these factors to the cultures results in reproducible enhancement of 
osteoclast formation and avoids the need to remove lymphocytes either by wash- 
ing or obtaining a lymphocyte-free population by selection of CD14-positive 
cells. Interestingly, the mechanism by which TGF-(3 and IL-4/IL-13 mediate their 
effects appears to be different. 

17. These concentrations give a final concentration of 25 ng/mL of M-CSF and 1 ng/mL 
of IL-4. 60 ng/mL of RANKL is added either at this time point (d 1) or can be 
added for the first time on d 4. Both protocols enhance osteoclast formation. 

18. IL-13 (0.1 ng/mL final concentration) exerts a similar effect on osteoclast forma- 
tion as IL-4. 

19. 180 [xL of medium are removed from these cultures in order to remove as much 
of the IL-4 as possible. The same volume of fresh medium is replaced containing 
25 ng/mL of M-CSF and 30 ng/mL of RANKL. If IL-4 is present in the culture 
medium in the absence of lymphocytes (lymphocytes are for the most part dead 
after 4 d) it exerts a powerful inhibitory effect on osteoclast formation 

20. There are major differences between the molecules that exert an osteoclast 
enhancing effect on murine precursors obtained from bone marrow (and spleen) 
and human precursors from peripheral blood. These differences may be due to 
the different sources of the osteoclast precursors. Alternatively, it may indicate 
that the molecules that regulate human and murine osteoclast formation are fun- 
damentally different. For example, TGF-(3 exerts no direct effect on human 
CD14-positive PBMNCs. In contrast, TGF-|3 is a powerful direct enhancer of 
murine osteoclast formation, exerting its effect directly on the murine bone mar- 
row M-CSF-dependent osteoclast precursor (11,12). 

21. If RANKL is added simultaneously with TGF-(3, there is no enhancement of 
osteoclast formation from unwashed PBMNCs. TGF-(3 is added only on d 1 of 
the culture and there is no additional benefit to adding TGF-|3 after day 1. We 
speculate that this is because TGF-|3 and IL-4/IL-13 abrogate the osteoclast- 
inhibitory effect of the lymphocytes, possibly blocking the effect of an osteoclast 
inhibitory factor produced by lymphocytes. Because lymphocytes are largely dead 
by d 4 in our cultures, the presence of TGF-|3 exerts no effect except when present 
at the beginning of the cultures. Unlike the effect of IL-4 and IL-13, TGF-|3 does 
not suppress osteoclast formation if maintained throughout the cultures. 

22. In phase II, even very high concentrations of M-CSF, up to 500 ng/mL, are not 
inhibitory to the process of osteoclast formation. In contrast, if high levels of M- 
CSF are added in phase I of the culture osteoclast formation is inhibited. 

23. The requirement for glucococorticoids (hydrocortisone/dexamethasone) in 
human osteoclast formation is complicated and apparently discrepant results are 
published. Dexamethasone, being more potent than hydrocortisone, is used at 10- 
8 M whereas hydrocortisone is used at 10-6 M. There is evidence that steroid 
hormones exert different effects on formation and resorption. Tobias and Cham- 

126 Flanagan and Massey 

bers showed that steroid hormones suppress bone resorption by the mature osteo- 
clast by inducing apoptosis (18) but there is also good evidence that steroid hor- 
mones induce osteoclast formation in human cultures in which either RANKL is 
provided endogenously by stromal cells or in human PBMNCs in which lympho- 
cytes are present (13). We have found, however, that steroid hormones are not 
necessary for human osteoclast formation (generated from CD34-positive pro- 
genitor cells, M-CSF-dependent precursors, or peripheral blood mononuclear 
cells) performed in the presence of soluble RANKL and that hydrocortisone sup- 
presses osteoclast formation in such cultures (2). Others have found that dex- 
amethasone enhances osteoclast formation from unwashed PBMNCs in which 
lymphocytes are present, in the presence of RANKL (19). Because, lympho- 
cytes suppress osteoclast formation formed from PBMNCs, these results sug- 
gest that steroid hormones abrogate the inhibitory effect of the lymphocytes. 
However, the reports are complicated by the results from Athanasou's group, 
who found that steroid hormones enhance osteoclast formation when lympho- 
cytes are absent in osteoclast-forming PBMNCs (personal communication). 
There is no clear explanation for these apparently discrepant results. However, 
one possibility is that the effect of steroid hormones is determined by the batch 
of serum used, as it is known that the concentration of steroid hormone in sera 
varies considerably. 

24. The purpose of phase I of the culture is to permit the osteoclast-supporting stro- 
mal (osteoblast/fibroblast) population of cells to expand. These are the cells that 
produce RANKL to promote osteoclast formation in phase II of the culture. 

25. Detailed time course experiments have been reported regarding the acquisition 
of the osteoclast phenotype (2,8). 

26. Osteoclast formation from this population of cells allows study of molecules that 
exert a direct action on the osteoclast precursor as bone marrow stromal cells and 
lymphocytes are removed from the cultures. It is clear that in the presence of 
cells other than osteoclast precursors, molecules can exert different effects from 
those seen if the precursors are present alone. For example, TGF-(3 exerts no 
effect on human osteoclast formation generated from CD14-positive cells 
whereas in a lymphocyte-rich environment TGF-|3 enhances osteoclast forma- 
tion from PBMNCs (5), but suppresses osteoclast formation in human bone mar- 
row cultures in which there is a stromal population (13). Another example is seen 
in the generation of murine osteoclasts; TGF-(3 enhances murine osteoclast for- 
mation by exerting a direct effect on the osteoclast precursor but suppresses 
osteoclast formation when osteoclasts are generated in the presence of bone mar- 
row stromal cells (11). 

27. It may be possible to plate the bone marrow precursors at a density of fewer than 
10 5 cells per well and to stop these experiments earlier that 14 d, but this has not 
been tested. 

28. DNase is added to bone marrow samples to prevent cell clumping (29). We have 
attempted to improve the osteoclast-forming potential of this culture system by 
treating the CD34-positive cells in various combinations of stem cell factor, flt3 

Human Osteoclasts from Marrow and Blood 127 

ligand, IL-3, IL-1 1, and TGF-(3 in the presence of M-CSF prior to the addition of 
and simultaneously with RANKL. However, these treatments did not enhance 
osteoclast formation over that generated in cultures treated with M-CSF and 
RANKL from d 1. 


1. Flanagan, A. M., Sarma, U., Steward, C. G., Vellodi, A., and Horton M. A. (2000) 
Study of the non-resorptive phenotype of osteoclast-like cells from patients with 
malignant osteopetrosis: a new approach to investigating pathogenesis. /. Bone 
Miner. Res. 15, 1-9. 

2. Lader, C. S., Scopes, J., Horton, M. A., and Flanagan, A. M. (2001) Generation of 
human osteoclasts in stromal cell-free and stromal cell-rich cultures: differences 
in osteoclast CD1 lc/CD18 integrin expression. Br. J. Haematol. Ill, 1210-1217. 

3. Sarma, U., Edwards, M., Motoyoshi, K., and Flanagan, A. M. (1998) 17(3-Estra- 
diol inhibits human osteoclast formation in vitro. /. Cell. Physiol. 175, 99-108. 

4. Lader, C. S. and Flanagan, A. M. (1998) Prostaglandin E2, interleukin la and 
tumor necrosis factor a increase human osteoclast formation and bone resorption 
in vitro. Endocrinology 139, 3157-3164. 

5. Massey, H. M., Scopes, J., Horton, M. A., and Flanagan, A. M. (2001) Transform- 
ing growth factor-(3 1 stimulates the osteoclast-forming potential of the 
haemopoietic precursor in peripheral blood cells in a lymphocyte-rich microenvi- 
ronment. Bone 28, 577-582. 

6. Scopes, J., Massey, H. M., Ebrahim, H., Horton, M. A., and Flanagan, A. M. 
(2001) Interleukin-4: bidirectional effects on human osteoclast formation. Bone 
29, 203-208. 

7. Fujikawa, Y., Quinn, J. M. W., Sabokbar, A., McGee, J. O. D., and Athanasou, N. 
A. (1996) The human osteoclast precursor circulates in the monocyte fraction. 
Endocrinology. 137,4058-4060. 

8. Massey, H. M. and Flanagan A. M. (1999) Human osteoclasts derive from CD14- 
positive monocytes. Br. J. Haematol. 106, 167-170. 

9. Kong, Y.-Y., Yoshida, H., Sarosi, I., et al. (1999) OPGL is a key regulator of 
osteoclastogenesis, lymphocyte development and lymph-node organogenesis. 
Nature 397, 315-323. 

10. Kong, Y. Y., Feige, U., Sarosi, L., et al. (1999) Activated T cells regulate bone 
loss and joint destruction in adjuvant arthritis through osteoprotegerin ligand. 
Nature 402, 304-308. 

11. Sells Galvin, R. J., Gatlin, C. L., Horn, J. W., and Fuson, T. R. (1999) TGF-J3 
enhances osteoclast differentiation in hematopoietic cell cultures stimulated with 
RANKL and M-CSF. Biochem. Biophys. Res. Commun. 265, 233-239. 

12. Fuller, K., Lean, J. M., Bayley, K. E., Wani, M. R., and Chambers, T. J. (2000) Arole 
for TGF(3(1) in osteoclast differentiation and survival. /. Cell Sci. 113, 2445-2453. 

13. Sarma, U. and Flanagan, A. M. (1996) Macrophage-colony stimulating factor (M- 
CSF) induces substantial osteoclast formation in human bone marrow cultures. 
Blood 88, 2531-2540. 

128 Flanagan and Massey 

14. Suda, T., Takahashi, N., Udagawa, N., Jimi, E., Gillespie, M. T., and Martin, T. J. 
(1999) Modulation of osteoclast differentiation and function by the new members 
of the tumor necrosis factor receptor and ligand families. Endocr. Rev. 20, 345-357. 

15. Lea, C. K., Sarma, U., and Flanagan, A. M. (1999) Macrophage colony stimulat- 
ing-factor transcripts are differentially regulated in rat bone-marrow by gender 
hormones. Endocrinology 140, 273-279. 

16. Flanagan, A. M. and Lader, C. S. (1998) Update on the biological effects of mac- 
rophage colony-stimulating factor. Curr. Opin. Hematol. 5, 181-185. 

17. Fuller, K., Wong, B., Fox, S., Choi, Y., and Chambers, T. C. (1998) TRANCE is 
necessary and sufficient for osteoblast-mediated activation of bone resorption in 
osteoclasts. /. Exp. Med. 188, 997-1001. 

18. Tobias, J. and Chambers, T. J. (1989) Glucocorticoids impair bone resorptive 
activity and viability of osteoclasts disaggregated from neonatal rat long bones. 
Endocrinology 125, 1290-1295. 

19. Matsukaki, K., Udagawa, N., Takahashi, N., et al. (1998) Osteoclast differentia- 
tion factor (ODF) induces osteoclast-like cell formation in human peripheral blood 
mononuclear cell cultures. Biochem. Biophys. Res. Commun. 246, 199-204. 

I ■ I I I 



Generating Murine Osteoclasts from Bone Marrow 

Naoyuki Takahashi, Nobuyuki Udagawa, Sakae Tanaka, 
and Tatsuo Suda 

1. Introduction 

Osteoclasts, the multinucleated giant cells that resorb bone, originate from 
hemopoietic cells of the monocyte-macrophage lineage (1,2). We have devel- 
oped a mouse bone marrow culture system, in which osteoclasts are formed in 
response to several bone-resorbing factors such as la, 25 -dihydroxy vitamin D 3 
[la,25-(OH) 2 D 3 ], parathyroid hormone (PTH), prostaglandin E 2 (PGE 2 ) and 
interleukin- 1 1 (IL-11) (2,3). We also developed a mouse coculture system of 
primary osteoblasts and hemopoietic cells to examine the regulatory mecha- 
nism of osteoclastogenesis (2,4). A series of experiments using the coculture 
system established the concept that osteoblasts/stromal cells have a key role in 
regulating osteoclast differentiation (2). Macrophage colony-stimulating fac- 
tor (M-CSF, also called CSF-1) produced by osteoblasts/stromal cells was 
shown to be an essential factor for differentiation of osteoclasts from osteo- 
clast progenitors (2,5). Recently, receptor activator of nuclear factor kB ligand 
(RANKL) was identified as another essential factor for osteoclastogenesis, 
which is expressed by osteoblasts/stromal cells in response to several bone- 
resorbing factors (6,7; see Note 1). Osteoclast precursors that possess RANK, 
a tumor necrosis factort (TNF) receptor family member, recognize RANKL 
through cell-cell interaction with osteoblasts/stromal cells, and differentiate 
into osteoclasts in the presence of M-CSF. Recent progress of molecular tech- 
nology allows us to introduce foreign genes into mature osteoclasts for modu- 
lating their functions. Adenoviral vectors are quite useful for introducing 
foreign genes into osteoclasts (8). We describe here the methods for osteoclast 
formation in mouse bone marrow cultures and for introduction of foreign genes 
into mature osteoclasts. 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


10/Takahashi/129-144/F1 129 I 2/26/03, 10:46 AM 

I ■ I I I 


130 Takahashi et al. 

2. Materials 

2. 1. Mice and Cell Lines 

1. ddY mice (see Note 2). 

2. Mouse bone marrow derived stromal cell lines, ST2 and MC3T3-G2/PA6 
(RIKEN Cell Bank, Tsukuba, Japan). 

3. Human embryonic kidney cell line 293 (American Type Culture Collection, 

Manassas, VA). 

2.2. Reagents 

1. Recombinant human M-CSF (Leukoprol; Kyowa Hakko Kogyo Co. Tokyo, 
Japan, or R &D systems, Minneapolis, MN) (see Note 3). 

2. Recombinant mouse TNF-a, and human IL-la (R&D Systems). 

3. la,25-(OH) 2 D 3 and PGE 2 (Wako Pure Chemical Industries, Ltd., Osaka, Japan). 

4. PTH (Peptide Institute, Inc., Osaka) and IL-1 1 (R&D Systems). 

5. Human osteoprotegerin (OPG) and a soluble form of human RANKL (Pepro Tech 
EC Ltd., London, UK). 

6. Synthetic analogue of eel calcitonin (Elcatonin, Asahi Chemical Industry Co. 
Tokyo, Japan). 

7. 125 I-labeled human calcitonin (Amersham Inc., Buckinghamshire, UK). 

8. NR-M2 emulsion (Konica Co., Tokyo). 

9. Rendol developer (Fuji Photo Film Co., Tokyo). 

10. Type I collagen gel solution (cell matrix type IA; Nitta Gelatin Co., Osaka) (see Note 4). 

11. Bacterial collagenase (Wako Pure Chemical Industries, Ltd.). 

12. Tissue culture plastics (Corning). 

13. oc-Modification of minimum essential medium a-MEM), Dulbecco's modified 
Eagle's medium (DMEM), and Ca 2+ - and Mg 2+ -free phosphate-buffered saline 
[PBS(-)] (Sigma Chemical Co., St. Louis, MO). 

14. Fetal bovine serum (FBS) (JRH Biosciences, Lenexa, KS or Gibco BRL, 
Gaithersburg, MY). 

15. Sterile instruments, syringes, and needles. 

16. Other chemicals and reagents are of analytical grade. 

2.3. Culture Media and Buffer Solutions 

1. a-MEM containing 10% FBS for cultures of mouse bone marrow cells. 

2. DMEM containing 10% FBS for cultures of 293 cells. 

3. PBS(-) for washing cells. 

4. a-MEM containing 0.2% bacterial collagenase for detachment of cells cultured 
on collagen gel coated dishes. 

5. Trypsin-EDTA solution: PBS(-) containing 0.05% trypsin and 0.5 mM EDTA 
for detachment of cells from culture plates. 

6. Pronase-EDTA solution: PBS(-) containing 0.001% pronase and 0.02% EDTA 
for removal of osteoblasts from cocultures. Pronase is dissolved in PBS(-) con- 
taining 0.02% EDTA just before use. 

10/Takahashi/129-144/F1 130 1 2/26/03, 10:46 AM 

I ■ I I I 


Murine Osteoclasts from Bone Marrow 131 

7. 0.1% Triton X-100 in PBS(-) for permeabilization of cells fixed with 3.7% form- 
aldehyde in PBS(-). 

8. Tartrate-resistant acid phosphatase (TRAP) staining solution: Five milligrams of 
naphthol AS-MX phosphate is dissolved in 0.5 mL of W,W-dimethyl formamide 
in a glass container. Thirty milligrams of fast red violet LB salt and 50 mL of 0. 1 M 
sodium acetate buffer, pH 5.0, containing 50 mM sodium tartrate are added to the 
mixture. This solution is made up fresh before use. Further details on TRAP stain- 
ing procedures can be found in Chapter 1 1 by van 't Hof and other chapters on 
osteoclast formation, this volume. 

9. Type I collagen mixture for preparing collagen gelcoated dishes: Type I collagen 
solution {see Subheading 3.3.), 5x cone. cc-MEM, and 200 mM N-2- 
hydroxyethylpiperazine-W-2-ethanesulfonic acid (HEPES) buffer, pH 7.4, con- 
taining 2.2% NaHC0 3 (7:2:1, by vol) are quickly mixed at 4°C just before use. 

10. 0.1 M cacodylate buffer, pH 7.4, containing 1% formaldehyde and 1% glutaral- 
dehyde for fixation of cells for autoradiography. 

3. Methods 

3. 1. Marrow Culture 

The mouse bone marrow culture system was developed for examining 
effects of bone-resorbing factors on osteoclast formation (3). Discovery of the 
RANKL-RANK interaction for osteoclastogenesis indicated that the growth 
of stromal cells is an essential step for osteoclast development in bone marrow 
cultures (6,7). 

1. Tibiae are removed aseptically from 7- to 9-wk-old male mice and the bone ends 
are cut off with scissors. The marrow cavities are flushed with 1 mL of a-MEM 
by injecting at one end of the bone using a sterile 27-gauge needle. 

2. Bone marrow cells are washed once with a-MEM, suspended in a-MEM con- 
taining 10% FBS, and cultured at 1.0 x 10 6 cells/0.5 mL/well in 24-well plates 
(Corning, Corning, NY) in a humidified atmosphere of 5% C0 2 . Cultures are fed 
every 2-3 d by replacing 0.4 mL of old medium with fresh medium see Note 5). 

3. Osteotropic factors such as 10~ 8 M la,25-(OH) 2 D 3 , 100 ng/mL of PTH, 10~ 6 M 
PGE 2 , and 10 ng/mL of IL-1 1 induce osteoclast formation in this marrow culture. 
These factors are usually added at the beginnning of culture and at each time of 
medium change. 

4. Cells are fixed, and stained for TRAP (a marker enzyme of osteoclasts) as 
described in section Subheading 3.5.1. 

TRAP-positive mononuclear cells appear on d 3-4 and multinucleated cells 
on d 4-5 in the presence of bone-resorbing factors. The number of TRAP- 
positive multinucleated cells reaches a maximum on d 6-8. TRAP-positive 
osteoclasts are formed only near the colonies of alkaline phosphatase (ALP)- 
positive osteoblasts in the culture treated with PTH (Fig. 1A,B; see Note 6). 
OPG completely inhibited the TRAP-positive cell formation induced by PTH 

10/Takahashi/129-144/F1 131 1 2/26/03, 10:46 AM 



■: *<*■;.. 

' . & * ■ • ; - • : ■ «■ 

*• * » **,**'■ ,' - 


Fig. 1. Enzyme histochemistry for TRAP and ALP in mouse bone marrow cultures. Bone marrow cells of ddYmice were cultured 
for7dwith lOOng/mL of (A) PTH, lOOng/mL of PTH plus 100 ng/mL of OPG (C), or 1 00 ng/mL of RANKL plus 50 ng/mL of M- 
CSF (E). Marrow cultures were then fixed and double-stained for TRAP and ALP. TRAP-positive cells appeared as red cells and 
ALP-positive cells as blue cells. (B), (D), and (F) show high power views of portions in (A), (C), and (E), respectively. Note that 
TRAP-positive cells formed in the culture are observed near or within the colonies of ALP-positive cells in the presence of PTH (B). 
In contrast, TRAP-positive cells are distributed uniformly on the culture dish in the presence of both RANKL and M-CSF (F). 
Adding OPG completely suppressed the formation of TRAP-positive cells induced by PTH (D). Scale bar = 200 \im. 


I ■ I I I 


Murine Osteoclasts from Bone Marrow 133 

in bone marrow cultures (Fig. 1C, D). Osteoclasts are also formed when mouse 
bone marrow cultures are treated with 50 ng/mL of M-CSF and 100 ng/mL of 
RANKL (Fig. IE, F). In this culture, osteoclasts are formed uniformly all over 
the culture dish. Cocultures of primary osteoblasts with bone marrow cells pro- 
duce more osteoclasts than bone marrow cultures alone do (4). A protocol for os- 
teoclast formation in coculture is given in Chapter 1 1 by van 't Hof, this volume. 

3.2. Bone Marrow Macrophage Culture 

Macrophages appearing in bone marrow cultures are the precursors of 
osteoclasts. We have modified the mouse bone marrow culture system to pre- 
pare highly purified osteoclast precursors (9). 

1. Bone marrow cells are treated with 100 ng/mL of M-CSF in aMEM containing 
10% FBS in 48-well plates (3 x 10 5 cells/0.5 mL/well) (see Note 7). 

2. Cells are cultured for 3 d, and nonadherent cells are completely removed from 
the culture by pipetting. Adherent cells strongly express macrophage specific 
antigens such as Mac-1, Moma-2, and F4/80. Therefore, adherent cells are called 
"M-CSF-dependent bone marrow macrophages (M-BMMt)))." Typically, 1 x 10 4 
M-BMMe|) are obtained when 1 x 10 5 bone marrow cells are cultured for 3 d in 
the presence of M-CSF. 

3. When M-BMMij) are further cultured with 100 ng/mL of RANKL and 100 ng/mL 
of M-CSF, TRAP-positive mononuclear and multinucleated cells are formed 
within 3 d (Fig. 2B) (see Note 7). 

3.3. Collagen Gel Culture 

Osteoclasts formed on plastic culture dishes are very difficult to detach by 
the treatment with either trypsin-EDTA or bacterial collagenase. To obtain 
functionally active osteoclasts formed in cocultures with osteoblasts, a col- 
lagen gel culture is recommended (10,11). 

1. A 10-cm culture dish (Corning) is coated with 4 mL of the type I collagen mix- 
ture at 4°C. The dish is put in a C0 2 incubator for 10 min to make the aqueous 
type I collagen gelatinous at 37°C. 

2. Primary osteoblasts (2 x 10 6 cells; see the chapter by Bakker and Klein-Nulend, 
this volume) and bone marrow cells (2 x 10 7 cells; see Subheading 3.1.) are 
cocultured on a collagen gel coated dish in 15 mL of a-MEM containing 10% 
FBS and 10~ 8 M la,25-(OH) 2 D 3 . The medium is changed every 2-3 d. 

3. After culture for 7 d, the dish is treated with 4 mL of 0.2% collagenase solution 
for 20 min at 37°C in a shaking water bath (60 cycles/min). The culture dishes 
are carefully placed on a sheet of aluminum foil put on the water surface of the 
water bath to maintain the sterile condition of the dishes. 

4. The cells released from the dish are collected by centrifugation at 250g for 5 min 
and suspended in 10 mL of a-MEM containing 10% FBS (the crude osteoclast 
preparation). Usually, 4 x 10 4 — 1 x 10 5 osteoclasts are recovered from a 10 cm 

10/Takahashi/129-144/F1 133 1 2/26/03, 10:46 AM 



Fig. 2. Effects of RANKL, mouse TNF-a , and IL-1 on TRAP-positive cell formation in M-BMMcj) cultures. Bone marrow cells of ddY 
mice were cultured with M-CSF for 3 d to prepare M-BMM<|>. M-BMM<|) were further cultured for 3 d without (A) or with either 100 ng/mL of 
RANKL (B), 20 ng/mL of mouse TNF-a (C), or 10 ng/mL of human IL-1 (D) in the presence of 100 ng/mL of M-CSF. Cells were then fixed 
and stained for TRAP. Scale bar = 100 u,m. 


I ■ I I I 


Murine Osteoclasts from Bone Marrow 135 

collagen gel coated dish, and the purity of osteoclasts is 2-3% in this crude prepa- 
5. The crude osteoclast preparation is used for biological and biochemical studies 
of osteoclasts. 

3.4. Purification of Osteoclasts Formed In Vitro 

Because the purity of osteoclasts in the crude osteoclast preparation is only 
2-3%, further purification is essential for biochemical studies of osteoclasts. 
Osteoclasts are easily purified from the crude osteoclast preparation placed on 
plastic dishes by treatment with pronase-EDTA solution (12,13). This proce- 
dure is identical to that described in the chapter by Coxon et al. to obtain pure 
mature rabbit osteoclasts. 

1. Ten milliliters of the crude osteoclast preparation is placed on a 10-cm culture 
dish (Corning) for 6-15 h in the presence of 10% FBS (Fig. 3A). 

2. Adherent cells are washed with oc-MEM, and treated with 8 mL of pronase- 
EDTA solution for 10 min. 

3 . Osteoblasts are then removed by gentle pipetting. More than 90% of the adherent cells 
on the dishes are TRAP-positive mononuclear and multinucleated cells (Fig. 3B). 

Using the purified osteoclast preparation, we have shown that osteoclasts 
possess phosphatidylinositol-3 kinase (14), rho p21 (15), and p60 c ~ src (16). We 
have also reported that osteoclasts express IL-1 receptors (17), TNF type I and 
type II receptors (18), and RANK (18), and they respond to cytokines through 
these receptors. 

It is difficult to obtain a highly enriched preparation of functionally active 
osteoclasts (see Chapter 7 by Coxon et at. and Chapter 6 by Collin-Osdoby et 
al., this volume). Using the disintegrin "echistatin," highly purified mono- 
nuclear and binuclear prefusion osteoclasts (pOCs) can be obtained from the 
coculture of mouse bone marrow cells and mouse osteoblastic MB 1 .8 cells treated 
with 10" 8 M la,25-(OH) 2 D 3 (19). The purity of pOCs in the preparation is about 
95% as determined by staining for TRAP. We have shown that pOCs them- 
selves fail to form resorption pits on dentine slices, but they form resorption 
pits in the presence of RANKL or IL-1 (17,18). 

3.5. Identification of Osteoclasts Formed In Vitro 

3.5. 1. TRAP staining (see Note 8) 

Cytochemical staining for TRAP is widely used for identifying osteoclasts 
in vivo and in vitro (2,7,13). 

1. Cells are fixed with 3.7% (v/v) formaldehyde in PBS(-) for 10 min, fixed again 
with ethanol-acetone (50:50, v/v) for 1 min, and incubated with the TRAP-stain- 
ing solution for 10 min at room temperature. 

10/Takahashi/129-144/F1 135 1 2/26/03, 10:46 AM 

I ■ I I I 


Takahashi et al. 



■■•■■■. '/.'-■ , 


Fig. 3. Purified TRAP-positive osteoclasts formed in cocultures of mouse osteo- 
blasts and bone marrow cells. Primary osteoblasts (2 x 10 6 cells) and bone marrow 
cells (2 x 10 7 cells) were cocultured for 7 d on a collagen gel coated dish. The dish was 
then treated with 0.2% collagenase solution to recover all the cells from the dish. The 
cells released from the dish were collected by centrifugation and suspended in 10 mL 
of ct-MEM containing 10% FBS (the crude osteoclast preparation). The crude osteo- 
clast preparation was placed on a 10-cm culture dish for 10 h in the presence of 10% 
FBS (A). The purity of osteoclasts in this crude preparation was only 2-3%. Adherent 
cells were washed with a-MEM, then treated for 10 min with 8 mL of pronase-EDT A 
solution. Osteoblasts were then removed by gentle pipetting. More than 90% of the 
adherent cells on the dish were TRAP-positive mononuclear and multinucleated cells 
(B). Scale bar = 200 jxm. 



TRAP-positive osteoclasts appear as red cells. An incubation period longer than 
10 min should be avoided, since cells other than osteoclasts become weakly posi- 
tive with time. 

After staining, cells are washed with distilled water, and TRAP-positive multi- 
nucleated cells having three or more nuclei are counted as osteoclasts under a 

3.5.2. Autoradiography for Calcitonin Receptors 

Osteoclasts have been shown to possess abundant calcitonin receptors 
(13,20). Expression of calcitonin receptors is one of the most reliable markers 



2/26/03, 10:46 AM 

I ■ I I I 


Murine Osteoclasts from Bone Marrow 137 

for identifying osteoclasts. Here we give a method for detection of calcitonin 
receptors by autoradiography, but immunocytochemical detection has been 
described also (see Note 9). 

1. For autoradiography with 125 I-labeled human calcitonin, cultures are prepared on 
plastic coverslips (<{>1 3.5 mm) placed in 24-well culture plates. 

2. Cells grown on the coverslips are washed with oc-MEM, and incubated with 0.2 nM 
125 I-calcitonin in the presence or absence of 200 nM unlabeled salmon calcitonin 
in a-MEM containing 0.1% bovine serum albumin (BSA) for 1 h at 20°C. 

3. Cells are washed three times with PBS(-) and fixed for 5 min with 0.1M cacody- 
late buffer, pH 7.4, containing 1% formaldehyde and 1% glutaraldehyde. 

4. The specimens are fixed again with ethanol-acetone for 1 min, and stained for 

5. The coverslips are then mounted on a glass slide, dipped in NR-M2 emulsion, 
and stored in a dark box at 4 C C. 

6. After incubation for 14 d, slides are developed in Rendol. Calcitonin receptors 
are identified by accumulation of dense grains due to 125 I-calcitonin binding, 
which disappear from the specimen when incubated with excess unlabeled calci- 

3.5.3. Pit Formation Assay 

When osteoclasts are placed on dentine slices, they form resorption pits 
within 24 h. A reliable pit formation assay was established using the crude 
osteoclast preparation and dentine slices (13,21). 

1. Dentine slices (§ 4 mm, 200 fxm thick) are prepared from ivory blocks using a 
band saw (BS-3000, Exakt, Germany) and a cutting punch. 

2. Dentine slices are cleaned by ultrasonication in distilled water, sterilized using 
70% ethanol, and dried under ultraviolet light. 

3. Dentine slices are placed in 96- well plates containing 0.1 mL/well of a-MEM 
with 10% FBS (a slice/well). A 0.1-mL aliquot of the crude osteoclast prepara- 
tion is transferred onto the slices. 

4. After a setting period of 60 min at 37°C, slices are removed, and placed onto 24- 
well plates containing a-MEM with 10% FBS (0.5 mL/slice/well). 

5. After incubation for 24-48 h, the medium is removed and 1 M NH 4 OH (1 mL/ 
well) is added to the wells for 30 min. 

6. Dentine slices are then cleaned by ultrasonication, stained with Mayer's hema- 
toxylin (Wako Pure Chemical Industries) for 35-45 s, and washed with distilled 

7. Resorption pits are clearly visualized with Mayer's hematoxylin under transmit- 
ted light. 

8. The number of resorption pits formed on dentine slices is counted under a light 
microscope. Alternatively, the resorbed area is measured using an image analysis 
system linked to the light microscope. For a detailed description of such a method 
using reflected light microscopy see Chapter 1 1 by van 't Hof, this volume. 

10/Takahashi/129-144/F1 137 I 2/26/03, 10:46 AM 

I ■ I I I 


138 Takahashi et al. 

3.6. Introduction of Foreign Genes into Osteoclasts (see Note 10) 

3.6.1. Preparation of Adenovirus Vector 

An adenoviral vector system is useful for introducing foreign genes into 
osteoclasts to study the regulation of osteoclast function (8). 

1. The recombinant adenovirus carrying a foreign gene under the control of the 
CAG (cytomegalovirus IE enhancer + chicken |3-actin promoter + rabbit |3-globin 
poly [A] signal) promoter is constructed by homologous recombination between 
the expression cosmid cassette and the parental virus genome in 293 cells (22). 

2. Titers of virus stocks are determined by an endpoint cytopathic effect assay with 
the following modifications. Fifty microliters of DME containing 10% FBS is 
dispensed into each well of a 96-well plate, then eight rows of threefold serial 
dilutions of the virus starting from 1(H dilutions are prepared. 

3. To each well 3 x 10 5 293 cells in 50 p,L of DMEM containing 10% FBS is added. 
The plate is incubated at 37°C in 5% C0 2 in air, and 50 \iL of DME containing 
10% FBS is added to each well every 3 d. 

4. After culture for 12 d, the endpoint of the cytopathic effect is determined by 
microscopy, and the 50% tissue culture infectious dose (TCID 50 ) is calculated. 
One TCID 50 approx corresponds to one plaque forming unit (PFU)/mL. 

5. The efficiency of infection is affected not only by the concentration of viruses and 
cells, but also by the ratio of viruses to cells, the multiplicity of infection (MOI). 
MOI is expressed as a measure of titer how many PFU are added to each cell. 

3.6.2. Infection of Adenovirus Carrying Foreign Genes to Osteoclasts 

1. Incubate mouse cocultures on d 4, when osteoclasts begin to appear, with a small 
amount of a-MEM containing the recombinant adenoviruses for 1 h at 37°C at a 
suitable MOI. We usually employ an MOI of 100 in our experiments. 

2. Wash the cells twice with PBS(-) and incubate further with a-MEM with 10% 
FBS at 37°C. Experiments are preformed 24 h after the infection. 

Using recombinant adenovirus carrying the lacZ gene, we have shown that 
the adenovirus carrying the lacZ gene can effectively infect osteoclasts with no 
apparent morphological changes or cellular toxicity (8) (Fig. 4). A high level 
of (3-galactosidase activity is observed in mouse osteoclasts infected with the 
recombinant adenovirus carrying the lacZ gene. The proportion of (3- 
galactosidase-positive osteoclasts increases in an MOI dependent manner, 
and > 80% of osteoclasts are positively stained at 100 MOI (Fig. 4). 

4. Notes 

1. In 1997, osteoprotegerin (OPG) and osteoclastogenesis inhibitory factor (OCIF), 
which inhibit osteoclast development in vivo and in vitro, respectively, were 
cloned independently by two different research groups (23,24). Incidentally, OPG 
and OCIF were the same protein molecule. OPG/OCIF is a member of the TNF 

10/Takahashi/129-144/F1 138 1 2/26/03, 10:46 AM 

I ■ I I I 

Murine Osteoclasts from Bone Marrow 



Fig. 4. An efficient adenovirus vector-mediated gene transfer to mouse osteoclasts 
as evidenced by cytochemical staining for |3-galactosidase activity. Mouse cocultures 
on d 4 were incubated with a small amount of a-MEM containing the recombinant 
adenovirus vector encoding the lacZ gene (AxCASLacZ) for 1 h at 37°C at an MOI of 
100. The cells were then washed twice with PBS(-) and further cultured for 2 d with a- 
MEM with 10% FBS at 37°C. Then, the cells were fixed for 10 min in 3.7% (v/v) 
formaldehyde in PBS(-) and washed in PBS(-). |3-Galactosidase activity was detected 
by immersing the cells into a staining solution. Cells expressing |3-galactosidase activity 
were stained as blue cells. More than 80% of multinucleated osteoclasts were transfected 
with the adenovirus carrying the lacZ gene at an MOI of 100. Scale bar = 100 jxm. 

receptor family, but it does not have a transmembrane domain, suggesting that 
OPG/OCIF functions as a circulating soluble factor. Subsequently, the cDNA 
encoding the binding molecule of OPG/OCIF was isolated from an expression 
library of the mouse stromal cell line ST2 and was named as osteoclast differen- 
tiation factor (ODF) (25). A ligand for OPG/OCIF was also cloned from an 
expression library of the mouse myelomonocytic cell line 32D, and was named 
as OPG ligand (OPGL) (26). OPGL was found to be identical to ODF. The bind- 
ing molecule of OPG/OCIF was a membrane-associated protein of the TNF ligand 
family. ODF/OPGL was also identical to TNF-related activation-induced 
cytokine (TRANCE) (27) and RANKL (28), which were independently cloned 
from mouse T cell hybridomas and mouse dendritic cells, respectively. RANK is 
the transmembrane receptor of ODF/OPGL/TRANCE/RANKL, which is 
expressed by osteoclast precursors and mature osteoclasts (Fig. 5). OPG/OCIF is 
a decoy receptor of ODF/OPGL/TRANCE/RANKL. Thus, ODF, OPGL, 
TRANCE, and RANKL are different names for the same protein, which is essen- 
tial for the development and function of osteoclasts (Fig. 5). The terms 
"RANKL," "RANK," and "OPG" are used in this chapter according to the guide- 



2/26/03, 10:46 AM 

I ■ I I I 


Takahashi et al. 

j M-CSFreceptoi (c-Fms) 





Osteoblast&troml cell 

(ja^5(QH)2D3, PTH, IL-1l) 


Fig. 5. Schematic representation of osteoclast differentiation regulated by osteo- 
blasts/stromal cells. Recent studies have revealed the role of new TNF receptor-ligand 
family members responsible for osteoclast formation. Osteotropic factors such as 
la,25-(OH) 2 D 3 , PTH, and IL-1 1 stimulate expression of RANKL in osteoblasts/stro- 
mal cells as a membrane-associated cytokine for induction of osteoclast differentia- 
tion in bone marrow cultures or in cocultures of osteoblasts and hemopoietic cells. 
OPG, a soluble decoy receptor of RANKL, is produced mainly by osteoblasts. OPG 
strongly inhibits the RANKL-induced differentiation of osteoclast precursors into 
osteoclasts. The terms "RANKL," "RANK," and "OPG" are used in this chapter 
according to the guideline of the American Society for Bone and Mineral Research 
(ASBMR) President's Committee on Nomenclature (29). 

line of the American Society for Bone and Mineral Research (ASBMR) 
President's Committee on Nomenclature (29). When spleen cells are treated with 
RANKL in the presence of M-CSF, osteoclasts are formed even in the absence of 
osteoblasts. OPG completely inhibits osteoclast formation in bone marrow cul- 
tures treated with osteotropic factors. RANKL and OPG are expressed mainly by 
stromal cells in bone marrow cultures. Osteotropic factors stimulate the expres- 
sion of RANKL and suppress OPG expression by stromal cells in mouse bone 
marrow cultures. 

Other strains of mice such as BALB/c, C57BL, and ICR can also be used for 
mouse osteoclast formation. 

Human M-CSF is effective in both human and murine cells, whereas murine M- 
CSF is effective in murine cells but not in human cells. 

4. Only this type of collagen (cell matrix type IA; Nitta Gelatin Co.) is suitable for 
this procedure. 

5. Because FBS is one of the important factors that affect osteoclast formation, care- 
ful batch testing of FBS is recommended. 





2/26/03, 10:46 AM 

I ■ I I I 


Murine Osteoclasts from Bone Marrow 141 

6. Osteoclast development proceeds within a local microenvironment of bone. 
This process can be reproduced ex vivo in a coculture of mouse calvarial osteo- 
blasts and hemopoietic cells. Some mouse stromal cell lines such as MC3T3- 
G2/PA6 and ST2 are capable of supporting osteoclastogenesis when cultured 
with mouse spleen cells (2,30). In such cocultures, osteoclasts are formed in 
response to various osteotropic factors including lrx,25-(OH) 2 D 3 , PTH, PGE 2 
and IL-1 1 . Cell-to-cell contact between osteoblasts/stromal cells and osteoclast 
progenitors is required to induce osteoclastogenesis (2,7). Subsequent experi- 
ments have established that the target cells of osteotropic factors for inducing 
osteoclast formation in vitro are osteoblasts/stromal cells (2,7). In bone mar- 
row culture, stromal cells present in bone marrow support osteoclast formation 
from the progenitors in response to osteotropic factors. Therefore, the growth 
of stromal cells is one of the determinants for osteoclast formation in bone 
marrow cultures (2,7). 

7. Treatment of bone marrow cells with a high concentration of M-CSF (100 ng/mL) 
for 3 d stimulates the proliferation of macrophages without growth of stromal 
cells. ALP-positive cells are seldom observed in the M-BMMc)) preparation. In 
the absence of M-CSF, most of the M-BMM<|) rapidly die within 3 d. No TRAP- 
positive cells are formed even in the presence of RANKL, when M-CSF is not 
added to the culture. Mouse TNF-a (20-100 ng/mL) also stimulates formation of 
osteoclasts from M-BMM(|) in the presence of M-CSF (Fig. 2C), but human TNF- 
a shows only weak activity in inducing TRAP-positive cell formation from M- 
BMMc)) even at a higher concentration (100 ng/mL). Osteotropic hormones and 
cytokines including la,25-(OH) 2 D 3 , PTH, PGE 2 and IL-1 (Fig. 2D) fail to induce 
osteoclast formation in M-BMMe|) cultures. 

8. Various alternatives to the TRAP staining protocol given here are described in 
the other chapters on osteoclasts in this volume. All are equally useable. 

9. Recently, Quinn et al. (31) developed a method for immunostaining of murine 
calcitonin receptors using polyclonal antibodies against rat calcitonin receptors. 
Osteoclasts (both mononuclear and multinucleated) formed in the coculture of 
mouse bone marrow cells and osteoblasts were specifically immunostained by 
the antibodies. 

10. To investigate the molecular mechanism of osteoclast function, it is necessary to 
modulate gene expression in osteoclasts by introducing foreign genes into the 
cells. Adenovirus vectors have several advantages (8). First, these vectors are 
capable of infecting a variety of terminally differentiated cells including neu- 
rons, hepatocytes, and osteoclasts. Second, recombinant adenovirus can be 
amplified easily to a very high titer in vitro. Third, adenovirus infection to 
the cells has been reported to require the interaction of the RGD sequence in 
the penton base of the virus with the cell surface of osteoclasts. We have 
successfully introduced such foreign genes into osteoclasts as epidermal growth 
factor receptor (8), C-terminus Src family kinase (csk) (32), dominant negative 
ras (33), constitutively active MEK1 (MAP kinase kinase) (33), dominant nega- 
tive IkB kinase 2, and constitutively active IkB kinase 2 (33). 

10/Takahashi/129-144/F1 141 I 2/26/03, 10:46 AM 

I ■ I I I 


142 Takahashi et al. 


The authors thank Drs. Kanami Itoh, Koji Suda, and Xiao Tong Li (Depart- 
ment of Biochemistry, School of Dentistry Showa University) for providing 


1. Roodman, G. D. (1996) Advances in bone biology: the osteoclast. Endocr. Rev. 
17, 308-332. 

2. Suda, T., Takahashi, N., and Martin, T. J. (1992) Modulation of osteoclast differ- 
entiation. Endocr. Rev. 13, 66-80. 

3. Takahashi, N., Yamana, H., Yoshiki, S., et al. (1988) Osteoclast-like cell forma- 
tion and its regulation by osteotropic hormones in mouse bone marrow cultures. 
Endocrinology 122, 1373-1382. 

4. Takahashi, N., Akatsu, T., Udagawa, N., et al. (1988) Osteoblastic cells are 
involved in osteoclast formation. Endocrinology 123, 2600-2602. 

5. Yoshida, H., Hayashi, S., Kunisada, T., et al. (1990) The murine mutation osteo- 
petrosis is in the coding region of the macrophage colony stimulating factor gene. 
Nature 345, 442-444. 

6. Yasuda, H,, Shima, N., Nakagawa, N., et al. (1999) A novel molecular mecha- 
nism modulating osteoclast differentiation and function. Bone 25, 109-113. 

7. Takahashi, N., Udagawa, N., and Suda, T. (1999) A new member of TNF ligand 
family, ODF/RANKL/TRANCE/OPGL, regulates osteoclast differentiation and 
function. Biochem. Biophys. Res. Commun. 256, 449^155. 

8. Tanaka, S., Takahashi, T., Takayanagi, H., et al. (1998) Modulation of osteoclast 
function by adenovirus vector-induced epidermal growth factor receptor. /. Bone 
Miner. Res. 13, 1714-1720. 

9. Kobayashi, K., Takahashi, N., Jimi, E., et al. (2000) Tumor necrosis factor a 
stimulates osteoclast differentiation by a mechanism independent of the ODF/ 
RANKL-RANK interaction. /. Exp. Med. 191, 275-286. 

10. Akatsu, T., Tamura, T., Takahashi, N., et al. (1992) Preparation and characteriza- 
tion of a mouse multinucleated cell population. /. Bone Miner. Res. 7, 1297-1 306. 

11. Suda, T., Nakamura, I., Jimi, E., and Takahashi, N. (1997) Regulation of osteo- 
clast function. /. Bone Miner. Res. 12, 869-879. 

12. Jimi, E., Ikebe, T., Takahashi, N., Hirata, N., Suda, T., and Koga, T. (1996) 
Interleukin- 1 13 activates an NF-KB-like factor in osteoclast-like cells. J. Biol. 
Chem. 271, 4605-4608. 

13. Suda, T., Jimi, E., Nakamura, I., and Takahashi, N. (1997) Role of la,25- 
dihydroxy vitamin D 3 in osteoclast differentiation and function. Methods Enzymol. 
282, 223-235. 

14. Nakamura, I., Takahashi, N., Sasaki, T., et al. (1995) Wortmannin, a specific in- 
hibitor of phosphatidylinositol-3 kinase, blocks osteoclastic bone resorption. 
FEBSLett. 361,79-84. 

10/Takahashi/129-144/F1 142 I 2/26/03, 10:46 AM 

I ■ I I I 


Murine Osteoclasts from Bone Marrow 143 

15. Zhang, D., Udagawa, N., Nakamura, I., et al. (1995) The small GTP-binding pro- 
tein, rho p21, is involved in bone resorption by regulating cytoskeletal organiza- 
tion in osteoclasts. /. Cell Sci. 108, 2285-2292. 

16. Tanaka, S., Takahashi, N., Udagawa, N., et al. (1992) Osteoclasts express high 
levels of p60 r ■'"', preferentially on ruffled border membranes. FEBS Lett. 313, 

17. Jimi, E., Nakamura, I., Duong, L. T., et al. (1999) Interleukin 1 induces 
multinucleation and bone-resorbing activity of osteoclasts in the absence of 
osteoblasts/stromal cells. Exp. Cell Res. 247, 84-93. 

18. Jimi, E., Akiyama, S., Tsurukai, T., et al. (1999) Osteoclast differentiation factor 
acts as a multifunctional regulator in murine osteoclast differentiation and func- 
tion. /. Immunol. 163, 434-442. 

19. Wesolowski, G., Duong, L. T., Lakkakorpi, P. T., et al. (1995) Isolation and char- 
acterization of highly enriched, prefusion mouse osteoclastic cells. Exp. Cell Res. 
219, 679-686. 

20. Takahashi, N., Akatsu, T., Sasaki, T., et al. (1988) Induction of calcitonin recep- 
tors by 1 a, 25-dihydroxy vitamin D 3 in osteoclast-like multinucleated cells formed 
from mouse bone marrow cells. Endocrinology 123, 1504-1510. 

21. Tamura, T., Takahashi, N., Akatsu, T., et al. (1993) A new resorption assay with 
mouse osteoclast-like multinucleated cells formed in vitro. /. Bone Miner. Res. 8, 

22. Miyake, S., Makimura, M., Kanegae, Y., et al. (1996) Efficient generation of 
recombinant adenoviruses using adenovirus DNA-terminal protein complex and 
a cosmid bearing the full-length virus genome. Proc. Natl. Acad. Sci. USA 93, 

23. Simonet, W. S., Lacey, D. L., Dunstan, C. R., et al. (1997) Osteoprotegerin: a novel 
secreted protein involved in the regulation of bone density. Cell 89, 309-319. 

24. Tsuda, E., Goto, M., Mochizuki, S., et al. (1997) Isolation of a novel cytokine 
from human fibroblasts that specifically inhibits osteoclastogenesis. Biochem. 
Biophys. Res. Commun. 234, 137-142. 

25. Yasuda, H., Shima, N., Nakagawa, N., et al. (1998) Osteoclast differentiation fac- 
tor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is iden- 
tical to TRANCE/RANKL. Proc. Natl. Acad. Sci. USA 95, 3597-3602. 

26. Lacey, D. L., Timms, E., Tan, H. L., et al. (1998) Osteoprotegerin ligand is a 
cytokine that regulates osteoclast differentiation and activation. Cell 93, 165-176. 

27. Wong, B. R., Rho, J., Arron, J., et al. (1997) TRANCE is a novel ligand of the 
tumor necrosis factor receptor family that activates c-Jun N-terminal kinase in T 
cells. /. Biol. Chem. 272, 25190-25194. 

28. Anderson, D. M., Maraskovsky, E., Billingsley, W. L., et al. (1997) A homologue 
of the TNF receptor and its ligand enhance T-cell growth and dendritic-cell func- 
tion. Nature 390, 175-179. 

29. The American Society for Bone and Mineral Research President's Committee on 
Nomenclature (2000) Proposed standard nomenclature for new tumor necrosis 

10/Takahashi/129-144/F1 143 1 2/26/03, 10:46 AM 


I ■ I I I 

144 Takahashi et al. 

factor family members involved in the regulation of bone resorption. /. Bone 
Miner. Res. 15, 2293-2296. 

30. Udagawa, N., Takahashi, N., Akatsu, T., et al. (1989) The bone marrow-derived 
stromal cell lines MC3T3-G2/PA6 and ST2 support osteoclast-like cell differen- 
tiation in cocultures with mouse spleen cells. Endocrinology 125, 1805-1813. 

31. Quinn, J. M., Morfis, M., Lam, M. H., etal. (1999) Calcitonin receptor antibodies 
in the identification of osteoclasts. Bone 25, 1-8. 

32. Miyazaki, T., Takayanagi, H., Isshiki, M., et al. (2000) In vitro and in vivo sup- 
pression of osteoclast function by adenovirus vector-induced csk gene. /. Bone 
Miner. Res. 15, 41-51. 

33. Miyazaki, T., Katagiri, H., Kanegae, Y., et al. (2000) Reciprocal role of ERK and NF- 
kB pathways in survival and activation of osteoclasts. /. Cell Biol. 148, 333-342. 

10/Takahashi/129-144/F1 144 I 2/26/03, 10:46 AM 


Osteoclast Formation 

in the Mouse Coculture Assay 

Robert J. van 'tHof 

1 . Introduction 

The murine coculture assay originally described by Takahashi et al. (1), was 
the first culture system developed that generated genuine, bone-resorbing 
osteoclasts. In this assay, osteoblasts are stimulated with 1,25-dihydro- 
xyvitamin D 3 (D3) to stimulate RANKL and macrophage colony-stimulating 
factor (M-CSF) expression. These factors then stimulate early osteoclast pre- 
cursors present in the spleen or bone marrow cell populations to differentiate 
into mature osteoclasts. At the end of the culture, osteoclasts can be identified 
by tartrate-resistant acid phosphatase (TRAP) staining, and, when the cultures 
are performed on dentine slices, resorption activity can be measured as well. 
Even though today it is possible to generate osteoclasts from bone marrow 
cells alone by treating the cultures with RANKL and M-CSF, the coculture 
system is still a useful model for studying osteoblast-osteoclast interactions. It 
has been widely used to study the origin of the osteoclast (2) and the effects of 
growth factors and drugs on osteoclast formation (3,4). In studies with osteo- 
petrotic mice, the coculture assay has been used to determine whether the un- 
derlying mechanism was due to a defect in the osteoblasts or in the osteoclast 
precursors (5). 

2. Materials 

2.1. General Reagents/Materials 

1. Sterile instruments (scissors, forceps). 

2. Sterile syringes and needles (19- and 25-gauge). 

3. Ficoll. 

4. Sterile Petri dishes. 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


146 van 't Hot 

5. Conical polypropylene centrifuge tubes. 

6. 25 -mL Centrifuge tubes. 

2.2. Tissue Culture Reagents 

1. Culture medium: a-Modification of minimum essential medium (a-MEM) 
supplemented with 10% fetal calf serum (FCS) and antibiotics. 

2. Hanks' balanced salt solution (HBSS). 

3. HBSS supplemented with 10% FCS. 

4. lOOOx stock l,25-(OH) 2 D 3 (10~ 5 M in ethanol, Sigma), further referred to as D3. 

5. Dentine slices: The dentine we use is elephant ivory, cut into slices of approx 200 pm 
thickness using a Buehler Isomet low-speed saw with a diamond wafering blade 
(series 15 HC). Out of these slices we punch discs that fit the wells of a 96- well 
plate, using a paper puncher (see Notes 1 and 2). 

2.3. TRAP Stain Reagents 

Solutions 2-5 are stable for months if kept protected from light in a refrigerator. 

1. Naphthol AS-BI-phosphate stock: 10 mg/mL of naphtol-AS-BI-phosphate in 
dimethyl formamide (Sigma). Stable ± 1 wk at 4°C. 

2. Veronal buffer: 1.17 g of anhydrous sodium acetate, 2.94 g Veronal (sodium 
barbiturate). Dissolve in 100 mL of distilled water. 

3. 0.1 N Acetate buffer, pH 5.2: 

a. Dissolve 0.82 g of anhydrous sodium acetate in 100 mL of distilled water. 

b. 0.6 mL of glacial acetic acid, make up to 100 mL with distilled water. 
Adjust the pH of solution a to pH 5.2 with solution b. 

4. 100 mM Tartrate: Dissolve 2.3 g of sodium tartrate in 100 mL of acetate buffer. 

5. Pararosanilin, acridinfrei: Add 1 g of Pararosanilin to 20 mL of distilled water 
and add 5 mL of concentrated hydrochloric acid; heat carefully for 15 min in a 
95 C C water bath while stirring and filter once the solution has cooled down. 

3. Methods 

3.1. Osteoblasts 

The assay starts with the isolation of the cell populations needed. Although 
some people have reported good results with osteoblast-like cell lines, such as 
ST2 cells, we have not been very successful with these and use primary osteo- 
blasts isolated from the calvaria of 2- to 3-d-old neonatal mice (see Note 5 in 
the chapter by Bakker/Klein-Nulend, this volume). The osteoblasts are plated on 
plastic or dentine 1 d before the addition of the bone marrow cells (see Note 1). 

3.2. Isolation of Bone Marrow Cells 

Although the assay was originally described using spleen cells (1; see Note 
3), we generally use bone marrow cells as the source of osteoclast progenitors. 
Furthermore, other people have successfully used certain hemopoietic stem 
cell lines, such as C2GM cells (6). 

The Coculture Assay 147 

1. Dissect the femurs and tibia out of two or three mice (3-6 mo old). 

2. Flush out the marrow using a 25-gauge needle and HBSS + 10% FCS. 

3. Get a single-cell suspension by squeezing the cell suspension through needles of 
decreasing size (start with 19-gauge, end with 25-gauge). 

4. Remove red blood cells by Ficoll density centrifugation (600g, 25 min, brake off). 

5. Harvest the bone marrow cells from the interface and wash once in HBSS. 

6. Resuspend in 1 mL of culture medium. 

7. Keep the cells on ice until use (but try to place the cells into culture as soon as 

3.3. Setting Up the Coculture 

The optimal number of bone marrow cells and osteoblasts per well may 
vary, as we have found that the optimal number depends on the mouse strain 
used. For the MF1 mouse strain we use, the following seeding densities give 
optimal numbers of osteoclasts. 

In a 96-well plate: 

1. On d 0, plate 8 x 10 3 osteoblasts in 100 iiL of medium + 10 nM D3 per well. 
Because of problems with evaporation of culture medium do not use the wells at 
the edges of the plate, but fill them with sterile water. 

2. On d 1, add 2 x 10 5 freshly isolated bone marrow cells in 50 \xh of medium/well. 
The medium should contain 10 nM D3. 

3. On day 6, refresh 50% of the medium as follows (see Note 4): 

a. Gently add 150 p,L of fresh medium + 20 nM D3 per well. 

b. Allow the cells to settle for ±15 min. 

c. Remove 150 |iL of medium from each well. 

4. On d 10, fix the cultures in formalin and perform a TRAP stain. 

The medium refresh needs to be done very carefully, because the confluent 
layer of osteoblasts can be quite easily disturbed, and come off (see Note 4). 
This would result in a total absence of osteoclasts. Usually the first osteoclasts 
and resorption pits appear on d 6 (see Note 5). Reasonable numbers of osteo- 
clasts (see Note 6) are present between d 7 and 10 (Fig. 1). 

3.4. TRAP Staining 

Osteoclasts express very high levels of the enzyme TRAP and can therefore 
be easily visualized by staining for this enzyme as follows (7; see Notes 7-10). 
As an alternative to the protocol described here, a staining kit from Sigma 
(387-A, Leukocyte Acid Phosphatase staining kit) can be used. This kit uses 
fast garnet as the dye, and this leads to a very dark purple stain. 

1. Rinse the cultures with PBS. 

2. Fix the cells for 5 min with 4% formaldehyde. 

3. Rinse with PBS. 

4. Prepare staining solution: 


van 't Hot 

Fig. 1 . End result of a coculture. (A) Multinucleated osteoclasts identified by TRAP 
staining. (B) Resorption pits visualized by reflected light microscopy. The resorption 
pits stand out as dark objects. (C) Actin rings in the osteoclasts, visualized by phalloi- 
din staining. 

Solution 1: 

In a glass container, add 150 u,L of naphthol AS-BI phosphate stock to 750 u,L 

Veronal buffer pH 10.1. Then add 0.9 mL of acetate buffer. Add 0.9 mL of 

acetate buffer with 100 mM tartrate. 
Solution 2: 

120 u,L of Pararosanilin. 

120,uLof4%NaNO 2 . 
Mix solutions 1 and 2, filter through a .45-nm filter and use immediately. 

5. incubate the cells for 30-60 min at 37°C with staining solution (50-100 p,L/well). 

6. Rinse with distilled water. 

7. Store in 70% ethanol. 

Osteoclasts and mononuclear osteoclast precursors should be visible as 
bright red stained cells {see Fig. 1A). 

The Coculture Assay 149 

3.5. Quantification of the Resorption Area 

After the osteoclasts have been stained and counted, the slices are cleaned 
and the resorption pits can be visualized either by staining with dyes such as 
toluidine blue or Coomassie blue, by scanning electron microscopy, or by 
reflected light microscopy. We routinely use reflected light microscopy, 
because it is easy to perform, the slices need only thorough cleaning and no 
staining, and the image obtained can be fairly easily quantified using image 
analysis. Because the slices need to be completely flat for the reflected light 
microscopy, we glue them to glass slides under pressure of a 0.5-kg metal 
weight. We use a Zeiss Axiolab reflected light microscope, fitted with a x 2.5 
lens, wide-field c-mount adapter, and Diagnostics Instruments Insight B/W 
large chip digital camera. This set up allows us to capture an entire bone slice 
in one image at sufficient resolution to identify and measure the resorption pits 
(Fig. IB). We developed our own image analysis software package using the 
Aphelion ActiveX image analysis toolkit from ADCIS (ADCIS SA, 
Herouville-Saint-Clair, France). The program prompts the user to focus the 
slice to be measured (Fig. 2A), captures the image, and identifies the dentine 
slice (Fig. 2B), so that any dark objects outside the dentine slice can be auto- 
matically removed. Next, the resorption pits are identified using gray level 
thresholding and selection by shape (Fig. 2C), and the resorption area is calcu- 
lated. The entire process takes 2-3 min per slice. 

4. Notes 

1. Ivory is preferred to cortical bone for this assay, because it is very even and does 
not contain osteocyte lacunae, which interfere with the identification of the re- 
sorption pits. 

2. As we use reflected light microscopy to visualize the resorption pits at the end of 
the culture, it is important that the surface of the slices is as smooth and shiny as 
possible. Therefore we polish the slices using brasso copper polish followed by 
cerium oxide polishing compound (Buehler) for fine polishing. Any remaining 
polish particles are removed by sonicating the dentine slices for ±15 min in 70% 
ethanol. The slices are stored in 70% ethanol until use. 

3. Spleen cells can be used as an alternative to bone marrow cells in the coculture. 
The advantage is that they are easier to isolate than bone marrow cells. However, 
we generally get more consistent results using bone marrow. To use spleen cells, 
dissect the spleens out of two young adult mice. Use a bent 19-gauge needle to 
press the cells out of the spleen. Thoroughly resuspend the cells, load onto Ficoll 
and proceed as described in Subheading 3.2., step 4 using similar cell numbers. 

4. The most common problem with the coculture is the osteoblast layer contracting 
and coming off the slice. This is usually due either to the plating density of the 
osteoblasts being too high or to the layer being disturbed during an over-vigorous 
medium refresh. Because of the last point, we do a 50% medium refresh and keep 

van 't Hot 

Fig. 2. Measurement of resorption pits. (Top left) An image of the dentine slice is 
captured using a digital camera. (Top right) The dentine slice is identified by the 
software. (Bottom) The resorption pits are detected using gray level thresholding. 



the number of medium refreshes as low as possible. Once half way through the 
culture is sufficient. Make sure that the tip of the pipet does not touch the osteo- 
blast layer. 

Any drugs or factors to be tested can be present during different parts of the 
coculture. To test effects on mature osteoclasts, drugs can be added during the 
last 2-3 d of the culture, whereas having the drugs present during the first 3-4 d 
gives an indication of effects on osteoclast precursors. 

The usual yield of osteoclasts should be between 150 and 300 osteoclasts per 
slice. If the numbers are substantially lower, this may be due to nonoptimal seed- 
ing densities. Although the seeding densities mentioned in the preceding work 
well for cells from MF1 mice, other mouse strains may need different densities. 
It should be noted that numbers of osteoblasts and bone marrow cells that are 

The Coculture Assay 151 

either too high or too low will both lead to a reduction of osteoclast numbers and 
a series of densities should be tested. 

7. In this murine assay, the most convenient procedure is the TRAP stain. However, 
it should be noted that in long term cultures, macrophage polykaryons become 
TRAP positive as well (8). These macrophage polykaryons can be distinguished 
from osteoclasts by their staining for the macrophage antigen F4/80. Further- 
more, the TRAP stain as described here works fine for murine osteoclasts. How- 
ever, when staining osteoclasts from different species, the concentration of 
tartrate may need to be changed. For human osteoclasts, for example, we use a 
final concentration of 100 mM tartrate. 

8. In some species, osteoclasts can be identified easily by staining for the vitronectin 
receptor (9). However, reagents for detection of vitronectin receptor in the mouse 
are not easily available. 

9. Osteoclasts can also be identified by the presence of calcitonin receptors (10). 
However, this procedure is too involved and time consuming for routine analy- 
sis, as it involves using radiolabeled calcitonin and autoradiography. 

10. Osteoclasts that are actively resorbing display an actin ring, and this can be visu- 
alized by staining the actin with phalloidin coupled to rhodamine (Molecular 
Probes or Sigma). Comparing total number of TRAP-positive osteoclasts with 
the number of cells displaying the actin ring gives an indication of the fraction of 
osteoclasts actively resorbing bone. 


1. Takahashi, N., Akatsu, T., Udagawa, N., et al. (1988) Osteoblastic cells are 
involved in osteoclast formation. Endocrinology 123, 2600-2602. 

2. Hagenaars, C. E., Kawilarang-De Haas, E. W., van der Kraan, A. A., Spooncer, E., 
Dexter, T. M., and Nijweide, P. J. (1991) Interleukin-3-dependent hematopoietic stem 
cell lines capable of osteoclast formation in vitro. /. Bone. Miner. Res. 6, 947-954. 

3. van 't Hof, R. J. and Ralston, S. H. (1997) Cytokine-induced nitric oxide inhibits 
bone resorption by inducing apoptosis of osteoclast progenitors and suppressing 
osteoclast activity. /. Bone. Miner. Res 12, 1797-1804. 

4. van 't Hof, R. J., Armour, K. J., Smith, L. M., et al. (2000) Requirement of the 
inducible nitric oxide synthase pathway for IL-1- induced osteoclastic bone 
resorption. Proc. Natl. Acad. Sci. USA 97, 7993-7998. 

5. Lowe, C, Yoneda, T., Boyce, B. F., Chen, H., Mundy, G. R., and Soriano, P. 

(1993) Osteopetrosis in Src-deficient mice is due to an autonomous defect of 
osteoclasts. Proc. Natl. Acad. Sci. USA 90, 4485-4489. 

6. De Grooth, R., Mieremet, R. H., Kawilarang-De Haas, E. W., and Nijweide, P. J. 

(1994) Murine macrophage precursor cell lines are unable to differentiate into 
osteoclasts: a possible implication for osteoclast ontogeny. Int. J. Exp. Pathol. 75, 

7. Barka, T. and Anderson, P. J. (1962) Histochemical method for acid phosphatase 
using hexazonium pararosanilin as coupler. /. Histochem. Cytochem. 10, 741-753. 

8. Modderman, W. E., Tuinenburg-Bol Raap, A. C, and Nijweide, P. J. (1991) Tar- 

152 van 't Hot 

trate-resistant acid phosphatase is not an exclusive marker for mouse osteoclasts 
in cell culture. Bone 12, 81-87. 
9. Horton, M. A., Taylor, M. L., Arnett, T. R., and Helfrich, M. H. (1991) Arg-Gly- 
Asp (RGD) peptides and the anti-vitronectin receptor antibody 23C6 inhibit den- 
tine resorption and cell spreading by osteoclasts. Exp. Cell Res. 195, 368-375. 
10. Nicholson, G. C, Horton, M. A., Sexton, P. M., et al. (1987) Calcitonin receptors 
of human osteoclastoma. Horm. Metab Res. 19, 585-589. 

I ■ I I I 



RANKL-Mediated Osteoclast Formation 
from Murine RAW 264.7 Cells 

Patricia Collin-Osdoby, Xuefeng Yu, Hong Zheng, 
and Philip Osdoby 

1. Introduction 

Osteoclasts (OCs) are the cells uniquely responsible for dissolving both the 
organic and inorganic components of bone during bone development and 
remodeling throughout life. These cells originate from hematopoietic precur- 
sors of the monocyte/macrophage lineage that are present both in the bone 
marrow and peripheral circulation, and their numbers and/or activity are fre- 
quently increased in a wide array of clinical disorders that are associated with 
excessive bone loss and affect millions of people (1). For many years, investi- 
gations into how OCs develop and function have been hampered by the consid- 
erable difficulties associated with isolating and culturing these normally rare 
cells. Whereas cell lines have frequently provided an invaluable research tool 
and are widely used to decipher mechanisms involved in osteoblast (OB) dif- 
ferentiation and bone formation, no immortalized cell lines for mature OCs 
exist and the few pre-OC cell lines that have been reported either do not 
undergo full OC differentiation (2,3) or involve coculture systems and cells 
that may not be readily available to all researchers (4-6). To compound the 
problem further, it has proven difficult or impossible until recently to generate 
reliably bone-resorptive OCs expressing mature OC characteristics from pri- 
mary bone marrow or circulating precursor cells in vitro. However, recent 
breakthroughs have now made the latter possible owing to the identification of 
the key signal, receptor activator of nuclear factor kB ligand (RANKL), that is 
responsible for the full development and activation of OCs both in vitro and in 
vivo (7-9). During OB development or in response to specific hormonal or 
local signals, RANKL becomes highly expressed on the surface of OB/stromal 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


12/Collin/153-166/F1 153 1 2/26/03, 10:46 AM 

I ■ I I I 


154 Collin-Osdoby et al. 

cells and directly interacts with a receptor, RANK, upregulated by macrophage 
colony-stimulating factor (M-CSF) on the surface of pre-OCs to stimulate their 
fusion, differentiation, and resorptive function. Many researchers now routinely 
form OCs in vitro through the exogenous addition of soluble recombinant 
RANKL (in combination with M-CSF to stimulate pre-OC proliferation, sur- 
vival, and RANK expression) to cultures of primary bone marrow cells or 
peripheral blood monocytes derived from various species (e.g., human, mouse, 
rat, rabbit, or chicken). However, such procedures still necessitate the isolation 
of primary precursor populations and in sufficient numbers to provide enough 
in vitro generated OCs for experimentation or characterization. 

In addition to primary cells, at least one pre-OC cell line, murine macroph- 
age RAW 264.7 cells, responds to RANKL stimulation in vitro to generate 
bone pit resorptive multinucleated OCs (RAW-OCs) with the hallmark charac- 
teristics expected for fully differentiated OCs (10-12). RAW cells have been 
extensively employed in macrophage studies for >20 yr and were originally 
established from the ascites of a tumor induced in a male mouse by intraperito- 
neal injection of Abelson leukemia virus (although RAW cells do not secrete 
detectable virus particles) (13). RAW cells express the c-fms receptor for 
M-CSF (14) as well as M-CSF (our unpublished observations), perhaps 
explaining why they also express high levels of RANK (10) and do not require 
M-CSF as a permissive factor in their RANKL-induced formation into RAW- 
OCs. RAW cells are increasingly being used for studies of OC differentiation 
and function. There are many advantages of this system over the generation of 
OCs from primary cell populations, including the: (1) ready access (making it 
unnecessary to schedule studies around when primary cells may become avail- 
able) and widespread availability of this cell line to most researchers; (2) easy 
culture and homogeneous nature of the pre-OC population (devoid of OBs, stro- 
mal, lymphocytes, or other cell types); (3) sensitive and very rapid RANKL- 
mediated formation of RAW-OCs (within d); (4) very large number of 
RAW-OCs that can be generated; and (5) high bone pit resorptive capability and 
expression of OC characteristics exhibited by RAW-OCs. In this chapter, we 
describe methods for the culture and RANKL-mediated differentiation of RAW 
cells into bone-resorptive RAW-OCs, the preparation of RAW-OC enriched 
populations by serum density gradient fractionation, and the culture and charac- 
terization of RAW-OCs. Such in vitro generated OCs can be analyzed using 
biochemical, immunological, physiological, molecular, or functional assays ac- 
cording to commonly employed procedures (see Chapter 6 by Collin-Osdoby, 
this volume). 

2. Materials 

2.1. Tissue Culture Medium, Solutions, and Supplies 

All media and solutions should be prepared with glass-distilled water. 

12/Collin/153-166/F1 154 I 2/26/03, 10:46 AM 

I ■ I I I 


RA W Osteoclast Formation 1 55 

1. Culture medium: Mix 90 mL of sterile Dulbecco's modified Eagle medium 
(DMEM) supplemented with 4 mM L-glutamine, 1.5 g/L of sodium bicarbonate, 
4.5 g/L of glucose, and 1.0 mM sodium pyruvate with 10 mL of fetal bovine 
serum (FBS, Gibco) and 1 mL of a lOOx stock of antibiotic/antimycotic (a/a, 
Gibco); store at 4°C and prewarm to 37°C for use with cells. 

2. Phosphate buffered saline (PBS): Add 9 g of NaCl, 0.385 g of KH 2 P0 4 , and 1 .25 g 
of K 2 HP0 4 per liter of water (final volume); adjust pH to 7.2 using 10 N NaOH. 

3. RANKL (Alexis Biochemicals, Peprotech, Calbiochem, R&D Systems, or 
homemade): Reconstitute and store as a concentrated stock solution (typically 
100 ng/mL in PBS) in small aliquots (-10-50 uL) at -80°C as recommended by 
the manufacturer, briefly thaw and dilute into culture medium (to 35 ng/mL final 
concentration for the homemade murine recombinant soluble RANKL) immedi- 
ately before use with RAW cells, and refreeze remaining RANKL (and aim to 
thaw individual vials no more than three times to retain optimal bioactivity). 

4. Moscona's high bicarbonate (MHB): Add 8 g of NaCl, 0.2 g of KC1, 50 mg of 
NaH 2 P0 4 , 1.0 g of NaHC0 3 , 2 g of dextrose, 10 mL of a/a, 990 mL of water, 
check that the pH is 7.2, and sterile filter. 

5. Hanks' balanced salt solution (HBSS, Gibco): Dissolve one packet in 990 mL of 
water, add 10 mL of a/a and 3.5 g of NaHC0 3 , check that the pH is 7.2, and 
sterile filter. 

6. Collagenase: Prepare 3% stock (3 g in 100 mL) solution in HBSS; store in aliquots 
(0.5-1.0 mL) at -20°C. 

7. Trypsin: Prepare 1% stock (1 g in 100 mL) solution in HBSS; store in aliquots 

8. Collagenase/trypsin digestion solution: Briefly thaw and add 71 \xL of 3% colla- 
genase solution and 141 uL of 1% trypsin solution to 3 mL of MHB (per dish) 
immediately before use with cells. 

9. Protease (EC, Sigma P-8811): Prepare 0.1% (100 mg in 100 mL) stock 
solution in PBS, store at 4°C for up to several months or in aliquots (0.5 mL) at -20°C 
for long-term storage. 

10. EDTA: Prepare 2% (2 g in 100 mL) stock solution (using EDTA sodium salt) in 
PBS, store at 4°C. 

1 1 . Protease-EDTA digestion solution: Briefly thaw and add 50 jxL of 0. 1 % protease 
solution and 50 jxL of 2% EDTA solution to 5 mL of PBS (per dish) immediately 
before use with cells. 

12. Supplies: Sterile bottles, flasks, and tissue culture dishes; rubber cell scrapers 
(Fisher); hemocytometer. 

2.2. Preparation of Devitalized Bone or Ivory Discs for Bone Pit Re- 
sorption Studies 

1 . Ivory is obtained either through donation from a local zoo or, in the United States, 
the Federal Department of Fish and Wildlife Services (or similar Department in 
other countries, see also the chapter by Nesbitt and Horton, this volume). Bovine 
cortical bone is obtained from a local slaughterhouse. Segments of ivory and 

12/Collin/153-166/F1 155 1 2/26/03, 10:46 AM 

I ■ I I I 


156 Collin-Osdoby et al. 

bovine cortical bone are thoroughly cleaned and washed (multiple HBSS and 
70% ethanol rinses), sliced into small chunks, and then reduced to rectangular 
0.4-mm thick sheets using a low-speed Isomet saw (Buehler, Lake Bluff, IL). 

2. The sheets are rinsed three times with 70% ethanol, incubated in 70% ethanol 
overnight, and then washed for several h in HBSS before circular discs are cut 
using a 5-mm paper punch. 

3. The discs are soaked repeatedly in 70% ethanol in sterile 50-mL tubes (alcohol 
changes can be gently poured off because the discs tend to stick to the side of the 
tube), and stored in 70% ethanol at -20°C. 

4. For experimental use, the required number of discs are removed from the tube 
using alcohol-presoaked tweezers (to maintain sterility) in a tissue culture hood, 
transferred to a fresh sterile 50-mL polypropylene tube, rinsed extensively by 
inversion and mild shaking at least three times with -40 mL sterile HBSS per 
wash, and the discs transferred using sterile tweezers into culture wells or dishes 
containing sterile HBSS for 3-24 h of preincubation in a tissue culture incubator 
prior to the plating of cells. HBSS is removed only immediately before the discs 
are to be used so that they do not dry prior to RAW cell or RAW-OC seeding. 

3. Methods 

3.1. RAW 264. 7 Cell Culture 

RAW 264.7 cells are obtained from the ATCC or similar cell repository. 
They represent a murine macrophage cell line that has the capability to be 
grown indefinitely as an OC precursor population or can be differentiated by 
treatment with RANKL into multinucleated bone pit resorptive OCs express- 
ing the hallmark characteristics of in vivo formed OCs (see Subheading 3.2.). 

All work should be performed in a sterile hood using sterile solutions and 

1. If starting from a frozen (liquid nitrogen) vial of RAW cells, quickly (<1 min) 
thaw the vial (e.g., in a 37°C water bath or by rapidly rubbing between palms), 
resuspend the cells in a small amount (-0.5 mL) of culture medium, and add the 
cell suspension to a T25 tissue culture flask. Increase the volume in the flask to 10 mL 
with additional culture medium and place into a tissue culture incubator (d 0). 

2. On d 3, withdraw the spent medium and refeed the cells with 10 mL of fresh 

3. Culture the cells until confluent (typically 4-5 d). 

4. To subculture the confluent RAW cells, withdraw the spent medium, add 10 mL 
of fresh medium to the flask, and scrape the cell layer into this fresh medium 
using a rubber scraper (see Note 1). 

5. Immediately add 0.01 mL of this cell suspension to 0.09 mL of fresh medium in 
a microcentrifuge tube, mix gently, and count the cells using a hemocytometer. 

6. Calculate the number of cells per milliliter. Plate the 10-mL RAW cell suspen- 
sion at 1.5 x 10 5 cells/cm 2 into tissue culture dishes of the desired size. Typically, 
one confluent T25 flask will provide a sufficient number of RAW cells to seed 

12/Collin/153-166/F1 156 1 2/26/03, 10:46 AM 

I ■ I I I 


RA W Osteoclast Formation 157 

two 100-mm dishes or 24- well dishes (see Notes 2 and 3). Increase these vol- 
umes with additional medium as needed to yield 8 mL per 100-mm dish or 0.5 (or 
1.0) mL per well of a 24-well dish, and then place the cells into a tissue culture 
7. To grow the RAW cells for an extended period of time, refeed the cultures every 
2-3 d and subculture when they reach confluency as in steps 4-6. 

3.2. RA W-OC Formation (see Note 4) 

This method is based on the published procedure of Hsu et al. (10). 

1. Culture RAW cells to confluency (see Subheading 3.1., steps 1-3). 

2. Subculture confluent RAW cells into 24-well dishes as described in Subheading 
3.1., steps 4-6 (see Notes 2-5). If the cells are to be used for cytochemical or 
immunological staining, replate the RAW cell suspension into 24-well dishes 
that contain a sterile glass coverslip in each well. If bone pit resorption is to be 
evaluated in parallel with OC development in the RAW cell cultures, replate the 
RAW cell suspension into 24-well dishes that contain two to four small discs of 
bone or ivory per well (with or without a glass coverslip under the discs). 

3. Immediately add soluble recombinant RANKL to the dishes at a final concentra- 
tion of 35 ng/mL to initiate OC development (d 0) and increase the volume in the 
wells to 0.5 (or 1.0) mL with additional culture medium (see Note 6). 

4. Culture to d 3. Briefly examine the cells under a microscope for evidence that 
RAW cells are beginning to fuse into multinucleated RAW-OCs. Withdraw the 
spent medium, and add 0.5 (or 1.0) mL of fresh medium containing 35 ng/mL of 
RANKL to the developing RAW-OC cell cultures. 

5. Culture until d 5 or 6 when many multinucleated RAW-OCs have formed but 
have not completely covered the dish (see Note 7). The d 5 or 6 RAW-OC popu- 
lations may be immediately fixed and used for cytochemical or immunological 
staining, harvested for biochemical or molecular studies, or analyzed for bone pit 
resorption (see Chapter 6). Alternatively, RAW-OCs can be purified further by 
serum gradient density fractionation (see Subheading 3.2.1.). 

3.2.1. Serum Gradient Purification of RAW-OC 

Because not all RAW cells fuse into multinucleated RAW-OCs by d 5 or 6, 
those that have can be purified from the remaining mononuclear cells using 
serum density gradient fractionation (see Note 8). This procedure is a modifi- 
cation of the one we routinely use to purify in vitro formed OC or OC-like cells 
from chick or human origin (17,18). Directions are provided for RAW-OCs 
formed on 100-mm tissue culture dishes. All steps are conducted at room tem- 
perature unless otherwise noted, and are performed in a sterile hood using ster- 
ile solutions and supplies. 

1. Remove the spent culture medium from two 100-mm dishes of d 5 or 6 RANKL- 
generated RAW-OCs. 

12/Collin/153-166/F1 157 I 2/26/03, 10:46 AM 

I ■ I I I 


158 Collin-Osdoby et al. 

2. Gently add 10 mL of MHB to each 100-mm dish to wash the cells. Remove and 
discard the washes. 

3. Repeat step 2 to wash the RAW-OCs twice more with MHB. 

4. Add 10 mL of MHB to each dish and place them into a tissue culture incubator at 
37°C for 15 min. 

5. Remove and discard the MHB solution from each dish. 

6. Add 5 mL of freshly prepared collagenase-trypsin digestion solution to each dish 
and incubate at 37°C for 5 min. 

7. Remove the dishes from the incubator and shake the plates gently by hand back 
and forth (e.g., slide the dish on a flat surface) for -30 sec to detach and loosen 
the interaction of cells with extracellular matrix produced by the cells. 

8. Completely remove the collagenase-trypsin solution containing the released ma- 
trix material from each dish and discard (see Note 9). 

9. Gently wash the adherent cells on each dish by releasing 10 mL of PBS slowly 
against the side wall of the dish. Completely remove and discard the washes. 

10. Repeat step 9 to wash the cells on each dish with PBS twice more. 

11. Add 5 mL of protease-EDTA digestion solution to each dish. Incubate at 37°C 
for 10-15 min (see Note 10). 

12. Loosen the adherent cells on each dish by flushing the protease-EDTA incuba- 
tion solution with a pipet gently over the surface of the cell layer to free the cells 
(see Note 11). 

13. Transfer the cell suspensions from two 100-mm dishes into one 50-mL sterile 
centrifuge tube containing 1.0 mL of FBS (to inhibit further protease action). 

14. Centrifuge the cells at lOOg for 5 min. 

15. Remove and discard the supernatant. Gently resuspend the cell pellet in 15 mL of 
MHB by repeatedly drawing up and releasing from a pipet (not too vigorously; 
see Note 11). 

16. Prepare 16 mL of 70% FBS in MHB (1 1.2 mL of FBS plus 4.8 mL of MHB) in a 
50-mL of centrifuge tube, and 16 mL of 40% FBS in MHB (6.4 mL of FBS plus 
9.6 mL of MHB) in another 50-mL tube. 

17. Prepare an FBS gradient in a 50-mL round-bottom centrifuge tube. To do this, 
carefully pipet 15 mL of the 70% FBS-MHB solution (from step 16) into the 
bottom of the tube. Very slowly overlay this with 15 mL of the 40% FBS-MHB 
solution (from step 16), using a pipet held at a 45° angle against the side of the 
tube just above the 70% FBS layer and slowly releasing the 40% FBS solution in 
a thin stream so as not to deform the surface of the 70% FBS layer. 

18. Let the tube stand undisturbed for 30 min (at room temperature) to allow the 
larger multinucleated RAW-OCs to settle under normal gravity and penetrate the 
FBS layers (see Note 12). 

19. Carefully pipet off the top 17 mL which contains mononuclear cells, and place 
into a 50-mL tube. 

20. Then, remove the 16-mL middle layer, which contains primarily mononuclear cells 
and some small multinucleated RAW-OCs, and place into another 50-mL tube. 

2 1 . The bottom 12-mL fraction contains predominantly large multinucleated RAW-OCs. 

12/Collin/153-166/F1 158 1 2/26/03, 10:46 AM 

I ■ I I I 


RA W Osteoclast Formation 1 59 

22. Centrifuge the purified RAW-OC bottom fraction (and the other fractions if they 
are also to be cultured and/or analyzed) at lOOg for 5 min. 

23. Gently resuspend the RAW-OC pellet in culture medium, count an aliquot in a 
hemocytometer, and plate 1000-4000 cells per well of a 24-well dish. Typically, 
the purified RAW-OCs from two 100-mm dishes can be cultured in 2-10 wells of 
a 24-well dish (with 0.5-1.0 mL of medium per well) for 6-24 h (see Note 13). 
The top and middle fractions from the serum gradient fractionation are typically 
cultured in 20-40 wells and 15-30 wells of a 24-well dish, respectively. Alterna- 
tively, RAW-OCs (and the top and middle fractions, if desired) may be used 
immediately for analysis (see Subheading 3.2., step 5). 

Serum gradient fractionation routinely provides 4000-10,000 purified 
RAW-OCs from one 100-mm dish (this depends on the efficiency of one's 
technique and, more importantly, on the exact stage of RAW-OCs used to 
purify the cells; see Notes 5, 7, 12, and 13). In unfractionated RANKL-gener- 
ated RAW-OC cultures, multinucleated (more than three nuclei) RAW-OCs 
typically represent ~1% on a per cell basis and 25% on a per nuclear basis of 
the total cell population (Fig. 1, left panel). In contrast, serum gradient purified 
RAW-OCs (more than three nuclei) typically comprise 60-90% on a per cell 
basis and 96% on a per nuclear basis of the total cell population in the bottom 
serum fraction (Fig. 1, lower right panel). On average, RAW-OCs in this bot- 
tom serum fraction contain 15-30 nuclei per cell. 

3.3. Assay Techniques 

3.3.1. Phenotypic and Functional Characterization of RAW-OCs 

Standard protocols can be used to evaluate the morphological (light, scan- 
ning electron microscope), ultrastructural (transmission electron microscope), 
histochemical (general or enzymatic activity stains), or immunocytochemical 
staining (e.g., for OC developmental markers) characteristics of RAW cells 
representing pre-OCs and in vitro RANKL-formed RAW-OCs (see Chapter 6 
by Collin-Osdoby et al., this volume). Whereas untreated RAW cells do not 
stain for tartrate resistant acid phosphatase (TRAP) activity, a key marker and 
enzyme involved in OC bone pit resorption, RANKL-differentiated RAW cell 
cultures develop both TRAP+ mononuclear and multinucleated cells (Fig. 
2A,C). The RAW-OCs formed by cell fusion contain multiple nuclei clustered 
together and the cells may appear either spread out or partially elongated when 
cultured on plastic (Fig. 2A). RAW-OCs cultured on bone or ivory (either dur- 
ing RANKL development or following replating of the differentiated cells) 
frequently display a more compact and highly motile elongated shape with 
numerous pseudopodial extensions (Fig. 2C). Resorption pits formed by RAW- 
OCs are typified by clusters of multilobulated excavation cavities or long re- 
sorption tracks (which may also be multilobulated) adjacent to or underlying 

12/Collin/153-166/F1 159 1 2/26/03, 10:46 AM 



Cell suspension 
-40% FBS — — 
-70% FBS 

Top fraction 

Middle fraction 
-Bottom fraction 


■ • ■ . 
5, :> , '_<#$ 


Fig. 1. RANKL-mediated RAW-OC formation and serum gradient purification. (Left) RAW cells were cultured with 35 ng/mL of 
murine recombinant RANKL for 6 d and then subjected to serum gradient fractionation. A well cultured in parallel was fixed and stained 
for TRAP activity to show the proportion of mononuclear and multinucleated TRAP+ cells that arise by d 6 in RANKL differentiated 
RAW cell cultures. The cells were viewed using a light microscope and images were captured with a computer-linked digital camera. 
(Reduced from original magnification, xlOO.) (Right) The top, middle, and bottom fractions from the serum gradient fractionation were 
replated and cultured on plastic for several hours, after which the cells were fixed and stained for TRAP activity. (Upper right) The top 
fraction consists entirely of mononuclear cells, some of which are TRAP+ (in contrast to untreated RAW cells which are all TRAP-, not 
shown). (Reduced from original magnification, x200.) (Middle right) The middle fraction primarily contains mononuclear cells, a 


I ■ I I I 

RA W Osteoclast Formation 




. - r- ■ * »jt i ' . J 


.. ■ 









* '• » 









** H* . .*■-'■ * 
' - _X / -, 1 1. 

Fig. 2. RANKL-mediated RAW-OC or MA-OC formation and bone pit resorption. 
(A, C) RAW cells were cultured with 35 ng/mL of murine recombinant RANKL for 6 d 
on plastic (A) or ivory (C), and then fixed and stained for TRAP activity. Note the well 
spread morphology of RAW-OCs on plastic (A) compared with the more compact and 
motile phenotype of such cells actively engaged in bone resorption on ivory (C). Abun- 
dant resorption pits and tracks were evident that were frequently composed of con- 
necting excavation cavities. These represent periods of RAW-OC attachment and pit 
formation, followed by RAW-OC movement to an adjacent area of ivory for further 
resorption. (A, C reduced from original magnification, x200.) (B, D) Murine bone 
marrow cells were isolated and cultured at 5.6 x 10 5 cells per well of a 24- well dish 
(1.9 cm 2 per well) with 25 ng/mL of murine M-CSF and 35 ng/mL of murine RANKL 
for 6 d on plastic (C) or ivory (D), after which the cells (MA-OC) were fixed and 
stained for TRAP activity. Like RAW-OCs, the TRAP+ MA-OCs were well spread on 
plastic (B) and more compact on ivory (D). Resorption pits and tracks formed by MA- 
OCs (D) were indistinguishable from those formed by RAW-OCs (B). (B, D reduced 
from original magnification, x200.) 

Fig. 1. (continued) portion of which are TRAP+, and some small multinucleated RAW- 
OCs. (Reduced from magnification, x200.) (Lower right) The bottom fraction con- 
sists primarily of large multinucleated RAW-OCs, although a few mononuclear cells 
may still be present. (Reduced from original magnification, xlOO.) 



2/26/03, 10:46 AM 

I ■ I I I 


1 62 Collin-Osdoby et al. 

RAW-OCs actively engaged in bone resorption (Fig. 2C). Molecular, immu- 
nological, and/or biochemical analyses have shown that RAW-OCs express 
hallmark properties of OCs including TRAP, calcitonin receptor, cathepsin K, 
matrix metalloproteinase-9, integrin av(33, and c-src (10,18, our unpublished 
data). Both the phenotypic and functional characteristics of RANKL-differen- 
tiated RAW-OCs resemble those of in vivo formed isolated murine OCs or 
RANKL-differentiated OCs (MA-OCs) formed from murine bone marrow cells 
in the presence of M-CSF (Fig. 2B, D). Thus, like RAW-OCs, TRAP + MA- 
OCs exhibit a well spread morphology on plastic (Fig. 2B) and a more com- 
pact, motile phenotype on bone or ivory (Fig. 2D). Multilobulated resorption 
pits and tracks formed by MA-OCs (Fig. 2C) also resemble those formed by 
RAW-OCs, with well defined margins and deep resorption lacunae (Fig. 2D). 
Resorption pit formation by RAW-OCs, in the presence or absence of modula- 
tors, can be quantified as for chicken or human OCs (see Chapter 6). In addi- 
tion to these phenotypic and functional analyses, RAW-OCs provide sufficient 
material (protein, RNA, etc.) for investigation of gene or protein expression 
levels, receptors and signal transduction pathways, production of various fac- 
tors (cytokines, growth factors, arachidonic acid metabolites), or the release of 
other substances (free radicals, enzymatic activities) (see Note 14). 

4. Notes 

1. We find that a rubber-tipped cell scraper works best because it completely con- 
tacts the surface of the tissue culture flask (or dish) and causes the least cell 

2. In general, RAW cells should be subcultured at a ratio of 1:3 to 1:6. 

3. If more cells will be needed than are provided by a reasonable number of T25 
flasks, the confluent RAW cells can be subcultured into T75 flasks (at a 1:3 to 
1:6 ratio) and then grown to confluency. 

4. The number of RAW cell passages affects RANKL-mediated OC formation. In 
our hands, RAW-OCs will no longer form in response to RANKL stimulation 
once they have undergone 18-20 passages from the time that they were received 
from the ATCC repository. The reason for this is not yet clear, although other 
researchers have similarly noted that not all RAW 264.7 cell lines (or passages?) 
will form OCs after RANKL treatment. It is also possible that particular lots of 
FBS may differentially influence RANKL-mediated RAW-OC formation. There- 
fore, at least several different lots and sources of FBS should be tested if there are 
difficulties encountered in trying to form RAW-OCs. 

5. The density of RAW cells plated affects the rate and yield of RAW-OC develop- 
ment, as well as the subsequent analysis of RAW-OCs formed. We find that too 
low a cell density (100-500 cells/cm 2 ) delays RAW-OC formation and decreases 
the final yield. For most purposes (e.g., testing the effects of various agents on 
RAW-OC development), the plating density for RAW cells should be in the range 

12/Collin/153-166/F1 162 I 2/26/03, 10:46 AM 

I ■ I I I 


RA W Osteoclast Formation 1 63 

of 10 3 to 3 x 10 4 /cm 2 to facilitate counting or characterization of the RAW-OCs 
formed and still generate a sufficient number for analysis. If RAW-OCs are to be 
purified by serum density gradient fractionation, the initial plating density of 
RAW cells should be considerably higher (1.5 x 10 5 /cm 2 ) so that enough RAW- 
OCs are obtained following their purification (see Subheadings 3.1. and 3.2.). 
However, too high a density of RAW cells (4.5 to 7.5 x 10 5 / cm 2 ) inhibits RAW- 
OC formation. 

6. In our experience, the potency of recombinant soluble RANKL for inducing OC 
formation is largely dependent on its source (not only for RAW cells, but also for 
human monocyte, mouse bone marrow, or chicken bone marrow preparations). 
Thus, different commercial RANKL preparations vary significantly in the dose 
required, kinetics of OC formation, and final yield of bone pit resorptive OCs 
obtained. This is not strictly due to species compatibility issues because human 
or murine recombinant RANKL are similarly efficient for inducing OCs from 
murine RAW or bone marrow derived cells, chicken bone marrow cells, or human 
peripheral blood monocytes (although all but RAW cells require M-CSF 
costimulation). Although we have successfully used various commercial RANKL 
preparations, we now routinely prepare larger quantities of soluble recombinant 
mouse RANKL in our own lab that exhibits high osteoclastogenic activity with 
murine RAW or bone marrow cells, chicken bone marrow cells, or human mono- 
cytes. At 35 ng/mL, this mouse RANKL induces multinucleated TRAP+ cell 
formation that is first apparent on d 3-4 of RAW cell culture. Lower RANKL 
concentrations delay the kinetics and final yield of RAW-OC formation. Others 
have used recombinant RANKL preparations in the range of 20-100 ng/mL 
(depending on its source and bioactivity) to form multinucleated TRAP+ cells 
that usually first appear on d 3-4 of RAW cell culture (11,15,16). However, cer- 
tain recombinant mouse RANKL preparations appear to require an additional 
anti-RANKL antibody crosslinking step to induce osteoclastogenesis (19). There- 
fore, it is recommended that pilot studies be performed with each new source and 
preparation of recombinant RANKL to ascertain an appropriate dose to achieve 
the level of RAW-OC formation needed. 

7. In our model of RANKL-mediated RAW-OC development, TRAP+ cells first 
become apparent on d 2 of culture, and multinucleated TRAP+ cells appear on d 2>—\ 
of culture and nearly reach a peak on d 5-6 of culture. It is important to either use 
the cells or subject them to serum gradient purification at this point (and not wait 
another d until the full peak of RAW-OC formation has occurred) because if the 
cells become overconfluent and overfused, they die very rapidly (<24 h) thereafter. 

8. The mononuclear cells that remain in the RANKL-treated RAW cell cultures by 
d 5-6 can still fuse to form more RAW-OCs on further culture, even in the 
absence of any additional RANKL stimulation. Therefore, such cells represent 
an early stage in OC differentiation and are not equivalent to the original non- 
RANKL-treated RAW cells (in a number of characteristics). It is not advisable to 
simply culture the RANKL-treated RAW cell populations for any longer period 
of time than d 5-6 to encourage more of the mononuclear cells to fuse into multi- 

12/Collin/153-166/F1 163 1 2/26/03, 10:46 AM 

I ■ I I I 


164 Collin-Osdoby et al. 

nucleated RAW-OCs because continuing cell fusion (even one extra d) produces 
cells that are too large and fragile, and which undergo rapid and massive apoptosis 
in the cultures once such RAW-OCs have formed. 
9. It is important to remove this released extracellular matrix completely, here and 
in the subsequent washes, or the cells will reattach to it and become extremely 
difficult to detach. At this point, the cells can be seen to have begun to pull up 
from the dish. 

10. A rounding up from the dish becomes more obvious at this point for the RAW- 
OCs. The cells may appear somewhat shrunken, but should still appear bright 
and viable. 

11. Use only wide-bore pipets or tips for any work in isolating or manipulating OCs 
to avoid fragmenting these large multinucleated cells. Also, care should be taken 
to resuspend, mix, or vortex OC preparations gently and for as little time as nec- 

12. The extent of RANKL-mediated RAW-OC formation impacts on this step as 
well. Therefore, if the RAW-OCs formed are relatively small (<10 nuclei per 
cell), they will be unable to settle effectively into the 70% portion of the serum 
gradient. However, if the RAW-OCs are too large (>30 nuclei per cell), they will 
be too fragile, prone to break, and die too quickly so that few viable RAW-OCs 
will be recovered following the serum gradient purification. 

13. Even under controlled conditions of RANKL-mediated RAW-OC formation as 
discussed in this chapter, once such cells have formed they tend to apoptose very 
rapidly and the cells can be lost within 24 h if allowed to develop too long. Addi- 
tion of 10 ng/mL of IL-la to promote RAW-OCs survival on plastic slows the 
apoptotic process only slightly. Therefore, we typically use RAW-OCs formed in 
tissue culture dishes by d 5 or 6 for analysis within 6-24 h (e.g., staining, RNA or 
protein extraction, etc.). If RAW-OCs have been formed on bone or ivory, 
resorption pits are usually evident by d 4 and maximal by d 6; modulators can be 
added at appropriate times to observe stimulatory or inhibitory effects on resorp- 
tion. When RAW-OCs are purified by serum gradient fractionation and replated 
onto tissue culture dishes, their viability is usually extended for an additional 24 h. 
Alternatively, purified RAW-OCs can be replated onto bone or ivory (-800 cells 
per well of a 48-well dish containing one piece of ivory or bone) and cultured 
with 35 ng/mL of RANKL and 10 ng/mL of interleukin-la (IL-la), in the pres- 
ence or absence of other modulators, for 5-6 d to ascertain effects primarily on 
preformed RAW-OCs (although some additional RAW-OC development also 
occurs during this time period, as the 70% serum purified fraction still contains 
some mononuclear cells). Purified RAW-OCs typically do not exhibit pit forma- 
tion within the first 24 h after replating onto bone or ivory. 

14. Although many of the phenotypic and functional characteristics of RAW-OCs 
match those of RANKL-differentiated primary murine bone marrow-derived OCs 
or isolated in vivo formed murine OCs, this cannot automatically be assumed to 
be true for any particular property being evaluated. The most obvious difference 
is the requirement for M-CSF in RANKL-stimulated OC formation from bone 

12/Collin/153-166/F1 164 I 2/26/03, 10:46 AM 

I ■ I I I 


RA W Osteoclast Formation 1 65 

marrow cells (which have relatively low RANK prior to M-CSF exposure) but 
not for transformed RAW cells (which already make M-CSF and express high 
RANK levels). Therefore, it is important to consider that the property under study 
in the RAW-OC cell system may not necessarily reflect that of normal murine 
OCs. Because RAW cells are easier to obtain and culture than primary bone mar- 
row cells, represent a relatively homogeneous population of pre-OCs (deficient 
in osteoblasts, stromal cells, lymphocytes, etc.), and provide abundant material 
for study, they provide a highly valuable resource for rapidly and efficiently 
screening and determining mechanisms underlying OC-related processes. How- 
ever, we recommend that such studies are subsequently followed by at least a lim- 
ited number of experiments using primary murine OCs (either isolated or 
RANKL-generated in vitro) to confirm that these processes are likewise observed 
in normal murine OCs and are not unique to transformed RAW cells or RAW-OCs. 


This work was supported by NIH Grants AR32927, AG15435, and AR32087 
to P. O. 


1. Roodman, G. (1996) Advances in bone biology — the osteoclast. Endoc. Rev. 17, 

2. Mancini, L., Moradi-Bidhendi, N., Brandi, M., Perretti, M., and Maclntyre, I. (2000) 
Modulation of the effects of osteoprotegerin (OPG) ligand in a human leukemic cell 
line by OPG and calcitonin. Biochem. Biophys. Res. Commun. 279, 391-397. 

3. Nagai, M., Kyakumoto, S., and Sato, N. (2000) Cancer cells responsible for 
humoral hypercalcemia express mRNA encoding a secreted form of ODF/ 
TRANCE that induces osteoclast formation. Biochem. Biophys. Res. Commun. 
269, 532-536. 

4. Hentunen, T., Reddy, S., Boyce, B., et al. (1998) Immortalization of osteoclast 
precursors by targeting Bcl-XL and Simian virus 40 large T antigen to the osteo- 
clast lineage in transgenic mice. /. Clin. Invest. 102, 88-97. 

5. Chen, W. and Li, Y. (1998) Generation of mouse osteoclastogenic cell lines 
immortalized with SV40 large T antigen. /. Bone Miner. Res. 13, 1 1 12-1 123. 

6. Takeshita, S., Kaji, K., and Kudo, A. (2000) Identification and characterization of 
the new osteoclast progenitor with macrophage phenotypes being able to differ- 
entiate into mature osteoclasts. /. Bone Miner. Res. 15, 1477-1488. 

7. Takahashi, N., Udagawa, N., and Suda, T. (1999) A new member of tumor necro- 
sis factor ligand family, ODF/OPGL/TRANCE/RANKL, regulates osteoclast dif- 
ferentiation and function. Biochem. Biophys. Res. Commun. 256, 449-455. 

8. Chambers, T. (2000) Regulation of the differentiation and function of osteoclasts. 
J.Pathol. 192, 4-13. 

9. Schoppet, M., Preissner, K., and Hofbauer, L. (2002) RANK ligand and 
osteoprotegerin. Paracrine regulators of bone metabolism and vascular function. 
Arterioscler. Thromb. Vase. Biol. 22, 549-553. 

12/Collin/153-166/F1 165 1 2/26/03, 10:46 AM 

I ■ I I I 


166 Collin-Osdoby et al. 

10. Hsu, H., Lacey, D., Dunstan, C, et al. (1999) Tumor necrosis factor receptor fam- 
ily member RANK mediates osteoclast differentiation and activation induced by 
osteoprotegerin ligand. Proc. Natl. Acad. Sci. USA 96, 3540-3545. 

11. Yamamoto, A., Miyazaki, T., Kadono, Y., et al. (2002) Possible involvement of 
IkappaB kinase 2 and MKK7 in osteoclastogenesis induced by receptor activator 
of nuclear factor kappaB ligand. /. Bone Miner. Res. 17, 612-621. 

12. Mizukami, J., Takaesu, G., Akatsuka, H., et al. (2002) Receptor activator of NF- 
kappaB ligand (RANKL) activates TAK1 mitogen-activated protein kinase kinase 
through a signaling complex containing RANK, TAB2, and TRAF6. Mol. Cell. 
Biol. 22, 992-1000. 

13. Raschke, W., Baird, S., Ralph, P., and Nakoinz, I. (1978) Functional macrophage 
cell lines transformed by Abelson leukemia virus. Cell 15, 261-267. 

14. Shadduck, R., Waheed, A., Mangan, K., and Rosenfeld, C. (1993) Preparation of 
a monoclonal antibody directed against the receptor for murine colony-stimulat- 
ing factor-1. Exp. Hematol. 21, 515-520. 

15. Koseki, T., Gao, Y., Okahashi, N., et al. (2002) Role of TGF-beta family in 
osteoclastogenesis induced by RANKL. Cell Signal 14, 31-36. 

16. Shin, J., Kim, I., Lee, J., Koh, G., Lee, Z., and Kim, H. (2002) A novel zinc finger 
protein that inhibits osteoclastogenesis and the function of tumor necrosis factor 
receptor-associated factor 6. /. Biol. Chem. 277 ', 8346-8353. 

17. Sells-Galvin, R., Cullison, J, Avioli, L., and Osdoby, P. (1994) Influence of 
osteoclasts and osteoclast-like cells on osteoblast alkaline phosphatase activity 
and collagen synthesis. /. Bone Miner. Res. 9, 1167-1178. 

18. Yu, X., Collin-Osdoby, P., and Osdoby, P. (2003) SDF-1 increases recruitment of 
osteoclast precursors by upregulation of matrix metalloproteinase-9 activity. Con- 
nect. Tissue Res. 44 (Suppl. 1), 1-6. 

19. Cappellen, D., Luong-Nguyen, N., Bongiovanni, S., Grenet, O., Wanke, C., and 
Susa, M. (2002) Transcriptional program of mouse osteoclast differentiation gov- 
erned by the macrophage colony-stimulating factor and the ligand for the receptor 
activator of NF-kappa B. /. Biol. Chem. 211, 21,971-21,982. 

12/Collin/153-166/F1 166 1 2/26/03, 10:46 AM 

I ■ I I I 



Analysis of Osteoclast Function 
in Mouse Calvarial Cultures 

Michael J. Marshall and Marit N. Rowlands 

1. Introduction 

Mouse calvarial organ cultures have been used for many years to study the 
basic mechanisms by which osteoclastic bone resorption is regulated. The most 
obvious advantage of organ cultures over in vivo studies is the absence of con- 
founding factors such as hormonal and mechanical influences. Calvarial organ 
cultures have an advantage over isolated cell systems also, in that they pre- 
serve the interrelationships between the different cell types in the bone, and the 
relationship between these cells and bone matrix. Despite this, the mouse cal- 
varial organ culture system has been used about half as frequently between 
1995-2000 as during 1990-1995, as shown by a text word search of scientific 
publications. The main reason for this is that osteoclasts, the cells responsible 
for resorbing bone, can now be isolated from bone tissue of several species 
including rat (1), chicken (2), mouse (3), rabbit (4), and human (5). More 
recently, techniques have been developed to generate osteoclasts in vitro (6,7) 
and then incubate these on bone slices to produce bone resorption. With these 
approaches, osteoclasts are defined by their ability to produce resorption pits 
on bone slices. However, this definition may not be sufficient. Other giant cells 
from nonosseous tissues can produce pits (8), and two distinct morphologies of 
multinucleate tartrate-resistant acid phosphatase (TRAP)-positive osteoclast- 
like cells have been described with differing resorptive capacities (9). Like the 
osteoclast, the foreign body giant cell is derived by fusion of monocytes, and if 
they share the same fusion mechanism then hybrid giant cells showing a range 
of phenotypes may be produced in cultures of bone marrow derived cells. The 
problem with studying osteoclasts in isolation is that this is not how they exist 
in vivo. Osteoclasts do not form in dead bone tissue, for instance, where a 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


13/Marshall/167-182/F1 169 1 2/26/03, 10:46 AM 

I ■ I I I 

1 70 Marshall and Rowlands 

fracture has damaged the blood supply (10). They arise by fusion of monocyte 
precursors that arrive in the blood capillaries (11). Osteoclast precursors must 
traverse the capillary endothelium and migrate to the bone surface, where they 
fuse to form multinucleated osteoclasts. They must be specifically activated to 
resorb bone in response to local or humoral stimulation (12) and they can be 
deactivated so that they detach from the bone surface (13). Osteoclasts are 
essential to the process whereby bone is adapted to local mechanical stress 
(14) and this can occur only via local mechanisms. Osteoclasts possess a spe- 
cific mechanism that recognizes a bone surface that is in need of remodeling 
(12). When recognition is triggered the resorption mechanism is initiated. This 
ensures that previously resorbed bone is not further resorbed so that resorption 
proceeds only to a fixed depth. Also, osteoclasts have a short life span, about 
4-12 d (15,16) and are responsive to apoptotic signals that bring about their 
orderly death (17). There is abundant evidence that each of these steps is under 
the control of other cell types, including endothelial cells (18,19), osteoblastic 
cells (20,21) and their descendants, the osteocyte and the bone lining cell (22), 
and, under inflammatory conditions, cells of the immune system (23). The 
clearest example of the control of osteoclasts by other cell types is the activa- 
tion of the receptor activator of nuclear factor-KB (RANK) receptor by RANK 
_^_ ligand (RANKL) expressed on osteoblasts and bone marrow stromal cells (24). 

In the absence of a resorptive stimulus, these cells default to the production of 
osteoprotegerin (OPG) which inhibits osteoclast activity and differentiation by 
binding and neutralizing RANKL (25). 

Bone organ cultures were pioneered by Dame Honor Fell (26) more than 50 yr 
ago and have been used by several investigators since (27-30). Various types 
of bones have been used, including fetal long bones showing endochondral 
ossification and neonatal calvaria showing intramembranous ossification. 
Nutritional requirements differ in these two models because the different 
mechanisms of ossification produce different structures. The planar nature of 
the calvaria means that diffusion path lengths are short whereas the cylindri- 
cal nature of the long bone can lead to long path lengths, particularly in post- 
natal bones. This tends to restrict the use of long bones in organ culture to 
late fetal development, and this in turn means that it is modeling rather than 
remodeling of bone that is studied. A large number of effectors of bone resorp- 
tion have been investigated including hormones, cytokines, growth factors, 
prostaglandins, drugs, and toxins (27-30) and the information obtained has in 
general been consistent with effects observed in vivo. 

2. Materials 
2.1. Dissection 

1. Scalpel (no. 3 with curved blade no. 15). 

13/Marshall/167-182/F1 170 1 2/26/03, 10:46 AM 

I ■ I I I 


Osteoclasts in Mouse Calvarial Cultures 1 71 

2. Fine scissors (Vannas-Tubingen spring scissors, 8.5 cm long with a 5-mm cut- 
ting edge). 

3. Scissors (Noyes spring scissors, 12 cm long with a 12-mm cutting edge). 

4. Watchmakers' forceps (Dumont no. 5, 11 cm long). 

5. Graefe forceps (10 cm long, curved, finely serrated) (Interfocus Ltd, Haverhill, UK). 

2.2. Microscopy and Image Analysis 

1. Dissecting microscope, Olympus with 0.7-4x zoom lens. 

2. Inverted microscope: Nikon Diaphot-TMD for phase contrast and fluorescence. 

3. Microscope: Leitz Dialux transmitted light and fluorescence microscope. 

4. Digital Camera: Nikon Coolpix 990. 

2.3. Media and Reagents 

All reagents should be made up sterile or sterilized by passing through a 
0.22-^im filter before use. 

1 . Radiolabeled calcium: Add 4 MBq/mL of [ 45 Ca] calcium chloride to 0.9% saline. 

2. Phosphate-buffered saline (PBS) with EDTA and heparin: Add 5 mM EDTA and 
10 U/mL of heparin to PBS. 

3. BGJb medium: Fitton Jackson modified BGJb medium supplemented with 10% 
heat inactivated fetal calf serum (FCS), 100 U/mL of benzylpenicillin, 100 u.g/mL 
of streptomycin (Glaxo), and 10~ 6 M indomethacin. Used for preculture of calva- 
rial explants (see Note 1). 

4. BGJb medium pH 6.8: Add 150 [iL of 1 M HC1 to BGJb medium containing 10% 
heat-inactivated FCS, 100 U/mL of benzylpenicillin, and 100 u.g/mL of strepto- 
mycin. Used for culture of calvarial explants (see Note 2). 

5. Medium 199: Medium 199 medium, with 25 mM W-2-hydroxyethylpiperazine- 
Af'-2-efhanesulfonic acid (HEPES), Hanks' balanced salt solution and 5% heat- 
inactivated FCS. 

6. Tissue culture plates: Selection of 12-, 24-, and 96-well plates. 

7. 1 M Hydrochloric acid to decalcify calvariae at the end of the culture period. 

8. Scintillation cocktail: FluoranSafe XE Scintran. 

2.4. TRAP Staining 

1. TRAP fixative: 95% Ethanol, 5% glacial acetic acid. 

2. TRAP staining reagent: 1 mM Naphthol AS-BI phosphate, 0.3 mg/mL of fast 
garnet GBC salt in 100 mM acetic acid and 26.8 mM L-(+)-tartaric acid. Adjust 
pH to 5.2 with NaOH and store at -20°C until use. 

3. Glutaraldehyde in HC1: 12.5% Glutaraldehyde in 1 M HC1. Make up fresh before use. 

4. PBS-EDTA: 10% EDTA in PBS, adjust to pH 6.6 with 1 M HC1. 

2.5. Integrin Staining 

1. Skim milk-Tris solution: 0.25% Skim milk in 0.1 M Tris-HCl, pH 7.4. 

2. Anti-(3 3 integrin antibody: Biotin-conjugated hamster IgG antimouse CD61 
(integrin |3 3 chain) monoclonal antibody (Cambridge Bioscience, Cambridge, UK). 

13/Marshall/167-182/F1 171 I 2/26/03, 10:46 AM 

I ■ I I I 


1 72 Marshall and Rowlands 

3. Control antibody: Biotin-conjugated hamster IgG monoclonal (anti-trinitrophe- 
nol, Cambridge Bioscience, UK). 

4. Avidin-peroxidase in skim milk-Tris: ExtrAvidin peroxidase (Sigma), 10 u.g/mL 
in skim milk-Tris solution. 

5. Peroxidase detection solution: 0.5 mg/mL of diaminobenzidene, 0.006% hydro- 
gen peroxide, 1 mM nickel chloride, 1 mM cobaltous chloride, 24 mM citric acid 
in 100 mM disodium hydrogen phosphate, pH 6.4. 

2.6. Bromodeoxyuridine Staining 

1 . Bromodeoxyuridine solution: 20 mM 5-Bromo-2'-deoxy uridine (BrdU) and 2 mM 
5-fluoro-2'-deoxyuridine (FdU) in 0.9% saline. Sterile filter before use. 

2. BGJb with BrdU and FdU. Add 50 \xM BrdU and 5 \iM FdU to BGJb medium. 

3. Methods 

3. 1. Calvarial Explant Culture 

1. Inject neonatal Balb/c neonatal mice (up to 2 d old) intraperitoneally with 10 jxL 
of radiolabeled calcium. Inject enough mice to give four or five bones per treat- 
ment group in the experiment (see Note 3). 

2. Euthanize the mice 5 d later by decapitation. 

3. After exposing the calvaria and pinning back the skin, pipet 100 \iL of PBS with 
EDTA and heparin onto the surface of the calvaria to prevent blood coagulation. 

4. Carefully dissect out the parietal bones (Fig. 1; see Note 4). 

5. Place the individual parietal bones into separate wells of a 12-well tissue culture 
plate containing 1 mL of medium 199 for 15 min. 

6. Replace with 1 mL of BGJb medium and preincubate for 24 h at 37°C in 5% C0 2 
and air. 

7. Replace the medium with 1 mL of BGJb medium, pH 6.8, containing test sub- 
stances or vehicle and incubate for 2 d at 37°C and in 5% C0 2 in air. 

8. Aspirate the conditioned medium from each well into an 1.5-mL Eppendorf tube. 

9. Removing the bone from each well using forceps, rinse briefly in PBS and put 
each bone into a scintillation vial containing 200 jxL of HC1 for 15 min. 

10. Add 4 mL of scintillation cocktail to the dissolved bone, cap the scintillation vial, 
and mix well. 

11. Add 200 uL conditioned medium from each well to a scintillation vial. 

12. Add 4 mL of scintillation cocktail to the medium, cap the scintillation vial, and 
mix well. 

13. Count the radioactivity in the bones and conditioned medium using a beta counter 
(see Note 5). 

14. Calculate the percentage resorption from each bone by subtracting the back- 
ground counts from the total counts released into the medium (i.e., medium counts 
x 5) divided by the sum of total counts in the medium and bone. 

15. Calculate the treated control ratio, by dividing the percentage calcium released 
from treated bones with that released by control bones (bones cultured in medium/ 
vehicle) (see Note 6). 

13/Marshall/167-182/F1 172 I 2/26/03, 10:46 AM 

I ■ I I I 

Osteoclasts in Mouse Calvarial Cultures 



frontal bone 
coronal suture 
sagittal suture 
lamboid suture 

parietal bones 

Fig. 1. The dorsal aspect of 7-d-old mouse calvaria and a diagram showing the 
suture lines. The initial incision is made with fine scissors posterior to and along the 
sagittal suture into the frontal bones. A cut is made along the full length of the lamboid 
suture, and then along the coronal suture which looks pale and is more difficult to see 
in this photograph. Flexing the parietal bone outwards, it hinges on the squamous 
suture, which is just out of sight on this photograph. Cut along the squamous suture to 
release the parietal bone. 

3.2. Quantitating TRAP-Positive Osteoclasts in Calvarial Bones 
and the Endocranial Membrane 

1. Eutahnize 7-d-old neonatal Balb/c mice by decapitation. 

2. Isolate the parietal bones and place in culture as described in Subheading 3.1., 
steps 1-7. 

3. Terminate the culture after 48 h by aspirating the medium and transferring the 
bones to individual wells of a 96-well plate filled with TRAP fixative. 

4. Transfer the bones, one at a time, to a Petri dish containing PBS and, with the aid 
of a dissecting microscope and watchmakers' fine forceps, strip the ectocranial 
membrane away and discard it (see Note 7). 



2/26/03, 10:46 AM 

I ■ I I I 


174 Marshall and Rowlands 

5. Carefully peel the endocranial membrane back to the sagittal suture. 

6. Transfer bones to individual wells of a 96-well plate containing 250 \xL of PBS. 

7. Stain the bone for TRAP activity by adding 1 00u,L of TRAP reagent to each well. 
Incubate for 15 min at 37°C. 

8. Wash each bone twice with 250 u,L of PBS. 

9. Fix and decalcify by adding 250 u,L of glutaraldehyde in HC1 to each well. 

10. Wash each bone three times with 250 u,L of PBS. 

11. With the aid of a dissecting microscope, mount the bones, with endocranial side 
uppermost and the endocranial membrane spread out in Aquamount on a glass 
microscope slide (Fig. 2a). 

12. Gently lower a coverslip onto the bone without disturbing the endocranial mem- 

13. Count TRAP -positive osteoclasts on each bone and membrane by transmitted 
light microscopy at xlOO magnification (see Note 8). 

3.3. Immunolocalizing the Integrin |3 3 Subunit in Calvarial Cultures 

1. Set up the calvarial cultures as described in Subheading 3.2., steps 1-8. 

2. Wash the calvariae by adding 1 mL of PBS to each well. 

3. Aspirate the PBS from each well. 

4. Replace with 250 u.L of PBS-EDTA and incubate at room temperature for 1 h. 

5. Aspirate and replace with 250p,L of fresh PBS-EDTA and incubate for a further 
30 min. 

6. Aspirate and add 250 u.L of skim milk-Tris solution to each well. Incubate for 10 min. 

7. Repeat step 5 a total of three times. 

8. Aspirate and add 25 \xL of skim milk-Tris solution containing 10 u.g/mL of anti- 
|3 3 antibody or control antibody to each test well. Incubate at room temperature 
for 1 h. 

9. Wash the bones three times with 200 u,L of skim milk-Tris solution. 

10. Add 25 \ih of avidin peroxidase in skim milk-PBS to each well; incubate for 1 h. 

1 1 . Wash each bone three times with 200 u,L of skim milk-Tris solution. 

12. Add 200 u,L of peroxidase detection solution to each well and incubate for 10 min 
at 37°C. 

13. Wash each bone in 250 u,L of PBS. 

14. With the aid of a dissecting microscope, mount the bones, with endocranial side 
uppermost and the endocranial membrane spread out in Aquamount on micro- 
scope slides. 

15. Gently lower a coverslip onto the bone without disturbing the endocranial membrane. 

16. Count |3 3 -positive (brown staining) osteoclasts in the bone and endocranial mem- 
brane by transmitted light microscopy (Fig. 3A,B; see Note 9). 

3.4. Analysis of Osteoclast Recruitment with BrdU 
in Mouse Calvarial Cultures 

1. Weigh 5-d-old neonatal mice and inject each mouse intraperitoneally with 
10 jxL/g of BrdU solution. 

13/Marshall/167-182/F1 174 I 2/26/03, 10:46 AM 

I ■ I I I 

Osteoclasts in Mouse Calvarial Cultures 



Fig. 2. (a) A freshly excised parietal bone showing the endocranial membrane 
peeled back to the sagittal suture and stained for TRAP. TRAP-positive osteoclasts are 
just visible on the bone surface on the right. (Reduced from original magnification, 
xl3.6.) (b) Irregular shaped TRAP-positive osteoclasts on the endocranial membrane 
of a parietal bone that has been incubated for 1 d in the presence of 10~ 6 M indometha- 
cin. (Reduced from original magnification, xl75.) (c) TRAP-positive osteoclasts on 
the parietal bone surface after incubation for 1 d in the presence of lO -6 M indometha- 
cin, then a further day in the presence of 10~ 6 M PGE 2 . Resorption lacunae are visible 
in dark outline behind some of these darkly stained osteoclasts. (Reduced from origi- 
nal magnification, xl75.) (d) TRAP activity tends to be concentrated at the leading 
edge of rapidly resorbing osteoclasts and is also deposited at the borders of the resorp- 
tion lacunae. Nuclei are visible in this osteoclast. (Reduced from original magnifica- 
tion, x470.) 

2. Euthanize by decapitation after 2 d. 

3. Set up calvarial cultures as described in Subheading 3.1., steps 1-6. 

4. Incubate in BGJb with BrdU for 24 h. 

5. Terminate the culture by transferring the bones to individual wells of 96-well 
plate filled with fixative. 

6. Transfer the bones, one at a time, to a Petri dish containing PBS and, with the aid 
of a dissecting microscope and watchmakers' fine forceps, strip the ectocranial 
membrane away and discard it. 



2/26/03, 10:46 AM 

I ■ I I I 


Marshall and Rowlands 


Fig. 3 (A) Immunolocalization of the |3 3 subunit of the integrin a v |3 3 . A parietal bone 
was incubated for 1 d in the presence of 10~ 6 M indomethacin then a further day in the 
presence of 10~ 6 M PGE 2 and then stained for the |3 3 subunit. Osteoclasts on the bone 
surface at about 12 h exhibit a pale brown concentric staining pattern which becomes 
more intense and crescent shaped at 24 h. (Reduced from original magnification, x222.) 
(B) At the end of a 1-d incubation of parietal bones in the presence of indomethacin, 
osteoclasts can be seen on the peeled endocranial membrane but they are very palely 
stained for the |3 3 subunit. (Reduced from original magnification, x222.) 

7. Carefully peel the endocranial membrane back to the sagittal suture. 

8. Transfer bones to individual wells of a 96-well plate containing PBS. 

9. Stain the bone for TRAP activity by adding 100 \xL of TRAP reagent to each 
well. Incubate for 15 min at 37 °C. 

10. Wash each bone twice with 250 \iL of PBS. 

11. Fix and decalcify by adding 250 [iL of HC1 with glutaraldehyde to each well. 

12. Wash each bone three times with 250 [xL of PBS. 

13. Add 200 u.L of 1 M HC1 to each well and incubate for 30 min at room temperature. 

14. Wash three times with 250 u,L of skim milk-Tris solution. 

15. Add 25 [iL of mouse monoclonal anti-BrdU antibody to each well. Incubate for 1 h 
at room temperature. 

16. Wash three times in skim milk-Tris solution. 

17. Add 25 [iL of sheep anti-mouse IgG peroxidase conjugate to each well and incu- 
bate for 1 h. 

18. With the aid of a dissecting microscope, mount the bones, with endocranial side 
uppermost and the endocranial membrane spread out in Aquamount on micro- 
scope slides. 

19. Gently lower a coverslip onto the bone without disturbing the endocranial membrane. 

20. Count BrdU-stained osteoclasts by transmitted light microscopy (see Note 10). 

4. Notes 

1. Role of preincubation: A 24-h preincubation period in BGJb medium with 
indomethacin is commonly employed to reduce spontaneous mineral mobiliza- 



2/26/03, 10:46 AM 

I ■ I I I 


Osteoclasts in Mouse Calvarial Cultures 1 77 

tion, which can occur at a high rate in some control cultures due to prostaglandin 
releases (30). It is thought that prostaglandin release occurs as the result of tissue 
damage during dissection, emphasizing the importance of a careful dissection 

2. Optimal pH for bone resorption: HC1 is added to culture medium during the 
experimental culture period because hydrogen ion concentration has a profound 
effect on the extent of bone resorption (27). At a pH above 7.2 bone resorption is 
inhibited whereas at pH 6.8 it is about maximal in (BGJb) medium (28). 

3. Number of bones per group: We typically use five bones for each set of culture 
conditions. To determine the extent of non-cell-mediated calcium release, it is 
possible to culture an additional group of five bones that have been frozen and 
thawed three times in liquid nitrogen. We do not do this routinely because in our 
experience calcium release is fairly constant at about 7.5%. If frozen bones are 
used, however, the percentage resorption for this group should be subtracted from 
the values obtained for test and control bones to obtain the cell-mediated calcium 

4. Use of half calvariae: Instead of using whole parietal bones it is possible to divide 
each parietal bone into two pieces and use half calvaria that contain the parietal 
bone and the frontal bone (31,32). These are often cultured at the interface of the 
medium and the gas phase held in position by stainless steel mesh supports. We 
and others (32) have found no advantage in supporting parietal bones on mesh 
supports over immersion in culture media. 

5. Alternative methods of assessing bone resorption: Calcium release may also be 
determined by the colorimetric analysis of calcium in the medium (33) instead of 
using radiolabeled calcium. This has the advantages of avoiding the use of radio- 
activity and the need to inject the mice. It may also result in a more precise mea- 
sure of resorption than that obtained by 45 Ca release because all the released 
calcium is measured and not just the radioactive calcium which may be incorpo- 
rated superficially within the bone. Assays are also available to measure products 
of bone degradation as an alternative to calcium measurements. They include 
measurements of the carboxy-terminal telopeptide fragments of rodent type I 
collagen, measurements of pyridinoline and deoxypyridinoline crosslinks, and 
measurements of hydroxyproline (34-38). These are more expensive than cal- 
cium measurements and it is unclear if they offer any benefits over calcium 

6. Analysis of paired bones: For a more sensitive detection of differences between 
test and control substances it is possible to calculate the ratio of the percentage 
calcium released from test to control wells using paired parietal bones from a 
single mouse. This approach takes advantage of the fact that within mouse varia- 
tion in parietal bone resorption is less than between mouse variation. 

7. Identifying the endocranial membrane: To be able identify the endocranial sur- 
face during dissection, it is helpful to cut a small diagonal section of bone from 
the left corner opposite the sagittal suture, with the concave endocranial side and 
the sagittal suture toward the operator. 

13/Marshall/167-182/F1 177 I 2/26/03, 10:46 AM 

I ■ I I I 


1 78 Marshall and Rowlands 

8. Distribution of TRAP-positive osteoclast in the bone and endocranial membrane: 
TRAP-positive osteoclasts can be counted using a conventional microscope by 
analyzing serial fields covering the entire bone and the endocranial membrane. 
Osteoclasts normally reside almost exclusively on the endocranial side of the 
bone in 7-d-old mice. After 1 d in culture in the presence of indomethacin, osteo- 
clasts tend to detach from the bone surface (39) and instead are seen attached to 
the endocranial membrane (13; Fig. 2B). This change in osteoclast distribution 
and osteoclastic bone resorption can be reversed if the bones are incubated with 
prostaglandin E 2 (PGE 2 ), 1,25-dihydroxy vitamin D 3 (l,25-(OH) 2 D 3 ) or parathy- 
roid hormone (PTH; Fig. 2C,D). 

9. After incubation with PGE 2 (3-3 staining is confined to osteoclasts on the bone 
surface Fig. 3A). At the end of the incubation with indomethacin, which causes 
osteoclasts to detach from bone and attach to the endocranial membrane, very 
few osteoclasts showed any |3 3 staining whether on the bone or on the endocra- 
nial membrane (Fig. 3B) (40). Other controls will be needed to discover the 
source of the problem if staining is seen in the presence of control antibody. 
Omitting the primary or the secondary antibody (or enzyme conjugate) will dis- 
tinguish between nonspecific binding of the primary antibody or enzyme conju- 
gate. Nonspecific binding may be due to inadequate blocking or too high a 
concentration of antibody(ies). Incubation with a range of antibody concentra- 
tions will indicate the best conditions for specific antigen detection. Incubation 
time and temperature can be varied to alter sensitivity. Omitting both the primary 
and the secondary antibody will demonstrate enzyme activity that is endogenous 
to the parietal bone. Endogenous peroxidase activity can be eradicated by incu- 
bation with 0.3% hydrogen peroxide in methanol for 30 min. Endogenous biotin 
can be blocked by preincubation with avidin. 

10. Studying the Kinetics of Osteoclast formation with BrdU. Multinucleate osteo- 
clasts arise by the fusion of monocyte precursors. To study the kinetics of osteo- 
clasts formation and osteoclast longevity we have developed a method for the 
detection of BrdU, an antigenic thymidine analogue, incorporated into the nuclei 
of monocytes that form osteoclasts on mouse parietal bones (41). The major ad- 
vantage of this method over [ 3 H]thymidine incorporation is that it does not require 
long exposure times for autoradiography. BrdU, along with FdU, can be injected 
into mice or added to culture medium without significant toxicity at the dose 
rates used. FdU increases the incorporation of BrdU into DNA by inhibiting en- 
dogenous thymidine synthesis (42). When BrdU is injected into mice, labeled 
osteoclast nuclei appear in the parietal bone 1 d later (Fig. 4). The population of 
labeled osteoclast nuclei decayed with a half-life of approx 1.3 d (41). When 
parietal bones were maintained in culture for up to 3 d in the presence of BrdU 
and FdU very few labeled osteoclast nuclei (fewer than four per bone) are seen. 
This suggests that there are very few osteoclast progenitors on mouse parietal 
bones of this age and therefore osteoclasts arise from postmitotic precursors that 
have arrived from the circulation. Other cell types in these parietal bones showed 
normal maintenance and labeling (41). BrdU labeling of osteoclasts has been 

13/Marshall/167-182/F1 178 1 2/26/03, 10:46 AM 

I ■ I I I 

Osteoclasts in Mouse Calvarial Cultures 



Fig. 4. BrdU incorporation into osteoclast nuclei. A 6-d-old mouse was injected 
with a saline solution containing BrdU and FdU, then 1 d later parietal bones were 
stained for TRAP and BrdU was immunolocalized. Three out of nine nuclei can be 
seen to be labeled with BrdU in this TRAP-positive osteoclast within a resorption 
lacuna. (Reduced from original magnification, x708.) 

used to examine the kinetics of osteoclast recruitment in mice in which bone 
resorption has been inhibited by the bis-phosphonate (3-amino-l- 
hydroxypropylidene)-l,l-bisphosphonate. In response to bisphosphonate, 
osteoclast recruitment is increased and they accumulated more nuclei than con- 
trols (43). Osteoclasts appeared to show degenerative morphology and had a 
shorter life-span (44). These observations are consistent with more recent obser- 
vations suggesting that bisphosphonates act directly on osteoclasts to inhibit key 
enzymes (45) that cause apoptosis (46). 


1. Chambers, T. J. and Magnus, C. J. (1982) Calcitonin alters behaviour of isolated 
osteoclasts. /. Pathol. 136, 27-39. 

2. Jones, S. J., Ali, N. N., and Boyde, A. (1986) Survival and resorptive activity of 
chick osteoclasts in culture. Anat. Embryol. 174, 265-275. 

3. Takada, Y., Kasuda, M., Hiura, K. et al. (1992) A simple method to access osteo- 
clast-mediated bone resorption using fractionated cells. Bone Miner. 17, 347-359. 

4. Chambers, T. J., Thomson, B. M., and Fuller, K. (1984) Effect of substrate com- 
position on bone resorption by rabbit osteoclasts. /. Cell Sci. 70, 61-71. 



2/26/03, 10:46 AM 

I ■ I I I 


180 Marshall and Rowlands 

5. Murrills, R. J., Shane, E., Lindsay, R., and Dempster, D. W. (1989) Bone resorp- 
tion by isolated human osteoclasts in vitro: effects of calcitonin. /. Bone Miner. 
Res. 4, 259-268. 

6. Takahashi, N., Akatsu, T., Uadagawa, N., et al. (1988) Osteoblastic cells are 
involved in osteoclast formation. Endocrinology 123, 2600-2602. 

7. Wani, M. R., Fuller, K., Kim, N. S., Choi, Y., and Chambers, T. (1999) Prostag- 
landin E 2 cooperates with TRANCE in osteoclast induction from hemopoietic pre- 
cursors: synergistic activation of differentiation, cell spreading and fusion. 
Endocrinology 140, 1927-1935. 

8. Athanasou, N. A. and Quinn, J. M. (1992) Bone resorption by macrophage 
polykaryons of a pilar tumor of scalp. Cancer 70, 469—475. 

9. Hata, K., Kukita, T., Akamine, A., Kukita, A., and Kurisu, K. (1992) Trypsinized 
osteoclast-like multinucleated cells formed in rat bone marrow cultures efficiently 
form resorption lacunae on dentine. Bone 13, 139-146. 

10. Burger, E. H., van der Meer, J. W. M., and Nijweide, P. J. (1984) Osteoclast 
formation from mononuclear phagocytes: role of bone forming cells. /. Cell Biol. 
99, 1901-1906. 

11. Fischman, D. A. and Hay, E. D. (1962) Origin of osteoclasts from mononuclear 
leucocytes in regenerating newt limbs. Anat. Rec. 143, 329-338. 

12. Suda, T., Nakamura, I., Jimi, E., and Takahashi, N. (1997) Regulation of osteo- 
clast function. /. Bone Miner. Res. 12, 869-879. 

13. Marshall, M. J., Holt, I., and Davie, M. W. J. (1996) Inhibition of prostaglandin 
synthesis leads to a change in adherence of mouse osteoclasts from bone to peri- 
osteum. Calcif. Tiss. Int. 59, 207-213. 

14. Hillam, R. A. and Skerry, T. M. (1995) Inhibition of bone resorption and stimula- 
tion of formation by mechanical loading of the modeling rat ulna in vivo. /. Bone 
Miner. Res. 10, 683-689. 

15. Loutit, J. F. and Townsend, K. M. S. (1982) Longevity of osteoclasts in radiation 
chimaeras of beige and osteopetrotic microphthalmic mice. Br. J. Exp. Pathol. 63, 

16. Marshall, M. J., Rees, J. A., Nisbet, N. W., and Wiseman, J. (1987) Reduced life 
span of the osteoclast in osteopetrotic (mi and mi dl ) mice. Bone Miner. 2, 1 15-124. 

17. Boyce, B. F., Hughes, D. E., and Wright, K. R. (1997) Methods for studying cell 
death in bone, in Methods in Bone Biology (Arnett, T.R. and Henderson, B., eds.), 
Chapman and Hall, London, pp. 127-148. 

18. Soskolne, W. A. (1979) The osteoclast-endothelium interface during bone resorp- 
tion in the femurs of young rabbits. Cell Tissue Res. 203, 487-492. 

19. Zaidi, M., Alam, A. S. M. T., Bax, B. E., et al. (1993) Role of the endothelial cell 
in osteoclast control: new perspectives. Bone 14, 97-102. 

20. McSheehy, P. M. J. and Chambers, T. J. (1987) 1,25-dihydroxy vitamin D 3 stimu- 
lates rat osteoblastic cells to release a soluble factor that increases osteoclastic 
bone resorption. /. Clin. Invest. 80, 425-429. 

21. Jimi, E., Nakamura, I., Amano, H., et al. (1996) Osteoclast function is activated 
by osteoblastic cells through a mechanism involving cell-to-cell contact. Endo- 
crinology 137, 2187-2190. 

13/Marshall/167-182/F1 180 1 2/26/03, 10:46 AM 

I ■ I I I 


Osteoclasts in Mouse Calvarial Cultures 181 

22. Lee, S., Harris, S. E., Roodman, et al. (2000) Chemotaxis: a potential mechanism 
whereby osteocytes target osteoclast precursors to bone. /. Bone Miner. Res. 15, 
(Suppl. 1) Abstr. F082, S208. 

23. Kong, Y., Felge, U., Sarosi, I., et al. (1999) Activated T cells regulate bone loss 
and joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature 
402, 304-309. 

24. Lacey, D. L., Timms, E., Tan, H.-L., et al. (1998) Osteoprotegerin ligand is a 
cytokine that regulates osteoclast differentiation and activation. Cell 93, 165-176. 

25. Yasuda, H., Shima, N., Nakagawa, N., et al. (1998) Osteoclast differentiation fac- 
tor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is iden- 
tical to TRANCE/RANKL. Proc. Natl. Acad. Sci. USA 95, 3597-3602. 

26. Fell, H. B. and Mellanby, E. (1952) The effect of hypervitaminosis A on embry- 
onic limb-bones cultivated in vitro. /. Physiol. 116, 320-349. 

27. Murrills, R. J., Dempster, D. W., and Arnett, T. R. (1997) Isolation and culture of 
osteoclasts and osteoclast resorption assays, in Methods in Bone Biology (Arnett, 
T. R and Henderson, B. eds.), Chapman and Hall, London, pp. 64-105. 

28. Biggers, J. D., Gwatkin, R. B. L., and Heyner, S. (1961) Growth of embryonic 
avian and mammalian tibiae on a relatively simple chemically defined medium. 
Exp. Cell Res. 25, 41-58. 

29. Lerner, U. H. and Gustafson, G. T. (1979) Inhibitory effect of dibutyryl cyclic 
AMP on the release of calcium, inorganic phosphate and lysosmal enzymes from 
calvarial bones cultured for 24 hours. Acta Endocrinol. 91, 730-742. 

30. Lerner, U. H. (1987) Modifications of the mouse calvarial technique improve 
responsiveness to stimulators of bone resorption. /. Bone Miner. Res. 2, 375-383. 

31. Reynolds, J. J. (1976) Organ cultures of bone: studies on the physiology and 
pathology of resorption, in Organ Culture in Biomedical Research (Balls, M. 
and Monnickendam, M., eds.), Cambridge University Press, Cambridge, UK, 
pp. 355-366. 

32. Ljunggren, O., Ransjo, M., and Lerner, U. H. (1991) In vitro studies on bone 
resorption in neonatal mouse cavariae using a modified dissection technique giv- 
ing four samples of bone from each calvaria. /. Bone Miner. Res. 6, 543-550. 

33. Meghji, S. Sandy, J.R., Scutt, A., Harvey, W., and Harris, M. (1988) Stimulation 
of bone resorption by lipoxygenase metabolites of arachidonic acid. Prostaglan- 
dins^, 139-149. 

34. Brown, S., Worsfold, M., and Sharp, C. (2001) Microplate assay for the measurement 
of hyroxyproline in acid-hydrolyzed tissue samples. BioTechniques 30, 38^-2. 

35. Black, D., Duncan, A., and Robins, S. P. (1988) Quantitative analysis of the 
pyridinium crosslinks of collagen in urine using ion-paired reversed-phase high- 
performance liquid chromatography. Analyt. Biochem. 169, 197-203. 

36. Gineyts, E., Garnero, P., and Delmas, P. D. (2001) Urinary excretion of 
glucosylgalactosyl pyridinoline: a specific marker of synovium degradation. 
Rheumatology 40, 315-323. 

37. Brady, J. D. and Robins, S. P. (2001) Structural characterization of pyrrolic 
crosslinks in collagen using biotinylated Erlich's reagent. /. Biol. Chem. 276, 

13/Marshall/167-182/F1 181 1 2/26/03, 10:46 AM 

I ■ I I I 


182 Marshall and Rowlands 

38. Robins, S. P. (1982) An enzyme-linked immunoassay for the collagen crosslink, 
pyridinoline. Biochem. J. 207, 617-620. 

39. Marshall, M. J., Holt, I., and Davie, M. W. J. (1995) The number of tartrate- 
resistant acid phosphatase-positive osteoclasts on neonatal mouse parietal bones 
is decreased when prostaglandin synthesis is inhibited and increased in response 
to prostaglandin E 2 , parathyroid hormone and 1,25 dihydroxy vitamin D 3 . Calcif. 
Tissue Int. 56,240-245. 

40. Holt, I. and Marshall, M. J. (1998) Integrin subunit (33 plays a crucial role in the 
movement of osteoclasts from the periosteum to the bone surface. /. Cell. Physiol. 
175, 1-9. 

41. Marshall, M. J. and Davie, M. W. J. (1991) An immunocytochemical method for 
studying the kinetics of osteoclast nuclei on intact mouse parietal bone. 
Histochem. J. 23, 402-408. 

42. Ellwart, J. and Dormer, P. (1985) Effect of 5-fluoro-2'-deoxyuridine (FdUrd) on 
5-bromo-2'-deoxyuridine (BrdUrd) incorporation into DNA measured with a 
monoclonal BrdUrd antibody and by the BrdUrd/Hoechst quenching effect. 
Cytometry 6, 513-520. 

43. Marshall, M. J., Holt, I., and Davie, M. W. J. (1993) Osteoclast recruitment in 
mice is stimulated by (3-amino-l-hydroxypropylidene)-l,l-bisphosphonate. 
Calcif. Tissue Int. 52, 21-25. 

44. Holt, I., Marshall, M. J., and Davie, M. W. J. (1994) Pamidronate stimulates 
recruitment and decreases longevity of osteoclast nuclei in mice. Semin. Arthrit. 
Rheum. 23, 263-264. 

45. Coxon, F. P., Helfrich, M. H., van 't Hot", R., Sebti, S. et al. (2000) Protein 
geranylgeranylation is required for osteoclast formation, function, and survival: 
inhibition by bisphosphonates and GGTI-298. /. Bone Miner. Res. 15, 1467-1476. 

46. Reszka, A. A., Halasy-Nagy, J. M., Masarachia, P. J., and Rodan, G. A. (1999) 
Bisphosphonates act directly on the osteoclast to induce caspase cleavage of mstl 
kinase during apoptosis. A link between inhibition of the mevalonate pathway 
and regulation of an apoptosis-promoting kinase. /. Biol. Chem. 21 A, 34,967- 

13/Marshall/167-182/F1 182 I 2/26/03, 10:46 AM 

I ■ I I I 



Assessing Bone Formation 

Using Mouse Calvarial Organ Cultures 

I. Ross Garrett 

1. Introduction 

1. 1. Historical Aspects 

The history of culturing bone explants goes back to the early 1920s, when 
Robinson reported that the enzyme alkaline phosphatase played an important 
role in bone mineralization based on studies of chick bone fragments. Subse- 
quently, Reynolds used bone explants to study collagen synthesis in bone (1) 
and to investigate the effects of a wide variety of other agents on bone turnover 
such as vitamin A (2), ascorbic acid (3), calcitonin (4,5), hydrocortisone (6) 
and bisphosphonates (7). Others have used rodent calvarial and long bone ex- 
plants to study the effects of cytokines on bone resorption (8-12) and bone 
formation (13,14). This chapter focuses on the use of cultured neonatal 
calvariae as an assay for agents with anabolic activity (13-19) (see Note 1). 

1.2. Overview of Assay 

The assay is essentially divided into three parts as shown in Fig. 1: (A) 
isolating the calvariae from 4-d-old mouse pups, (B) culturing the calvariae in 
media containing factors or compounds and, (C) measuring new bone forma- 
tion in treated calvariae by histological assessment of stained sections. The 
timelines for the assay are outlined in Figs. 2 and 3. For a 4-d assay (Fig. 2) the 
calvariae are dissected on day and incubated with compounds or factors over- 
night. On the following day (d 1), the calvariae are placed into fresh media and 
the cultures returned to the incubator and allowed to remain undisturbed until d 4, 
at which time the experiment is ended. For a 7-d assay (Fig. 3), the protocol is 
similar except that the media are replaced again on d 4 and the cultures incu- 
bated for another 3 d until d 7. The calvariae can be incubated for up to 2 wk 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


14/Garrett/183-198/F1 183 1 2/26/03, 10:46 AM 

I ■ I I I 



Murine neonatal calvaria 




Media /"" **"\ $rlJ 


Calvarial bone 

Bone lining cells 

Fig. 1. Overview of the calvarial assay (A) Dissection of the calvaria. (B) Incuba- 
tion in media on stainless steel grids. (C) Histological sections are cut to determine 
amount of new bone matrix formed. 












Fig. 2. Assay time line for 4-d assay. On d 0, calvariae are dissected and placed into 
media containing factors or compounds. On d 1, media are replaced with new media 
either with or without compounds and incubated. On d 4, calvarial bones are removed 
and histologically processed to determine the amount of new bone formation. 

with evidence of continued bone formation (see Note 2). Following process- 
ing, cutting, and staining of the sections from the treated calvariae the area of 
new bone formation is measured by using image analysis techniques. This 
measures the total area of bone and the amount of original bone remaining, the 
number of cells lining the bone tissue, as well as the thickness of a suture area. 



2/26/03, 10:46 AM 

In Vitro Bone Formation Assay 185 


















Fig. 3. Assay time line for 7-d assay. On d 0, calvaria are dissected and placed into 
media containing factors or compounds. On d 1, media are replaced with new media 
either with or without compounds and incubated. On d 4, media are replaced by fresh 
media with or without factors or compounds and incubated. On d 7, calvarial bones are 
removed and histologically processed to determine the amount of new bone formation. 

The dissection, culturing, histological processing, and measurement usually 
requires 10-14 d to complete and offers the ability to determine statistically 
new bone growth, changes in bone cell numbers, as well as a visual representa- 
tion of the bone formation occurring in the culture. 

2. Materials 

2. 1. Animals 

The assay is performed on bones dissected from 4-d-old mouse pups. These 
can be bred in house or purchased from an external supplier. We obtain the 
pups from pregnant Swiss white female mice, ordered to arrive on d 15 of 
gestation, but other strains of mice have been utilized for this assay with simi- 
lar results. The mice usually give birth on the 18— 19th day of gestation and 
produce litters of 10 or more pups. When the pups are 4 d old they are 
euthanized and the calvarial bones dissected out. 

2.2. Culture Media 

We use BGJ medium supplemented with 1 mg/mL of bovine serum albumin (Conn 
fraction V), 100 U/mL each of penicillin/streptomycin, and 0.292 mg/mL of glutamine. 
If using powdered medium, make up fresh with distilled water, add supplements, adjust 
to pH 7.2, and sterile filter before use. 

2.3. Preparation of Stainless Steel Grids 

1. Prepare the grids by cutting rectangular 1 x 1.5 cm pieces of stainless steel mesh 
from the sheet (we use stainless steel no. 30 mesh; Small Parts Inc. CX-60). 

2. Bend the ends of the mesh over using a sharp edge to make a bridge support (Fig. 
4) so that the dissected calvariae sit at the air-liquid interface when 1 mL of 
medium has been added to the tissue culture well (Fig. 5). After each assay, wash 
the grids in detergent and rinse in 10% nitric acid. Wash the cleaned grids three 
times in distilled water to remove any residual detergent and nitric acid. 

3. Dry the grids and heat sterilize in glass Petri dishes. Store sterile until use. 

I ■ I I I 




Fig. 4. Stainless steel grids measuring 1 cm x 1.5 cm are cut from mesh and both 
ends are bent to form a bridge to support the calvaria in culture. The mesh is bent at 
each end using a straight edge to make the bridge and adjusted to the height of 1 mL of 

Fig. 5. Image of a half neonatal murine calvaria in culture with concave or cranial 
side down on the stainless steel grid in 1 mL of medium. 

2.4. Staining Solutions 

1. Harris hematoxylin with acetic acid: Dissolve 25 g of ammonium aluminum sul- 
fate [A1NH 4 (S0 4 ) 2 12H 2 0] in 225 mL of dH 2 0. Dissolve 1.25 g of hematoxylin 
and 0.6 g of mercuric oxide in 15 mL of 95% ethanol and slowly add to the 
ammonium aluminum sulfate solution. Boil for 2 min and allow to cool. Store in 
the dark for at least 24 h. Immediately before use, filter the solution and add 8 mL 
of glacial acetic acid (see Note 3). 

2. 0.6% Eosin Y: Add 6 g of eosin Y to 900 mL of absolute ethanol. Stir to dissolve 
and adjust pH to between 4.6 and 5.0 with glacial acetic acid. 

3. 1% Phloxine B: Add 1 g of Phloxine B to 100 mL of dH 2 .0 

4. 2% Orange G: Add 2 g of Orange G, sodium salt to 100 mL of dH 2 0. 

5. Final eosin staining solution: Combine 6 mL of 2% Orange G with 6 mL of 1% 
Phloxine B and 238 mL of 0.6% eosin Y. 



2/26/03, 10:46 AM 

I ■ I I I 

In Vitro Bone Formation Assay 


3 ^ - 2 

Fig. 6. Procedure and steps for dissecting half neonatal murine calvaria from mice. 
The calvaria are dissected using the five different cuts. 


2.5. Fixation and Decalcification 

1. Fixation solution: 10% Buffered formalin. 

2. Decalcification solution: Prepare a solution of 14% EDTA in distilled water and 
adjust the pH to 7.2 by adding 9.5-10% NH 4 OH. 

3. Tissue processing: 75%, 95%, and 100% ethanol; xylene; paraffin wax. 

2.6 Histomorphometry 

1 . We analyze the sections using a Nikon E400 microscope to which is attached an 
Optronics three-chip color video camera. The output is fed to a 266-mHz Pentium 
PC computer using a Pro-Series 128 Capture Kit frame grabber board to obtain 
images of the sectioned calvaria. The area of new bone is then measured using 
Image Pro Plus (Media Cybernetics L.P., Del Mar, California). 

3. Methods 
3. 1. Dissection 

Dissect the calvariae free from the skull using fine dissection scissors and 
curved fine point forceps as illustrated in Fig. 6. 

1. Make sure the instruments used for dissection are clean and heat sterilized or 
autoclaved prior to dissection. The instruments must be kept sterile during the 
dissection procedure. 

2. Using the scissors, cut the calvaria along the midline or sagittal suture from the 
rear to the front past the coronal suture (cut 1). 

3. Make a second cut along the rear of the lambdoidal suture from the midline to a 
small capillary vessel above the ear (cut 2). 

4. Repeat this on the other side (cut 3). 



2/26/03, 10:46 AM 

I ■ I I I 


188 Garrett 

5. Dissect the calvariae free from the rest of the skull by cutting toward the front of 
the calvarial bones (cuts 4 and 5) and excising a half calvaria that contains both 
the frontal and parietal bones including the coronal suture. 

6. Once removed, carefully wash the half calvariae in a Petri dish containing media. 

3.2. Culture 

1. Place one sterile grid into each well of a 12-well tisue culture plate. 

2. Add 1 mL of culture medium to each well. 

3. Place one half calvaria in each well concave side down on the grids so that the 
calvaria is resting at the air-medium interface (Fig. 5). 

4. Incubate for 24 h at 37°C, in air-5% C0 2 in a humidified atmosphere. 

5. Using fine curved forceps, carefully transfer the grids, complete with the calva- 
rial bones, into a new 12-well plate containing media to which test substances 
have been added. Be careful not to disturb the calvariae during this process. Aim 
for at least four calvariae per treatment group. 

6. Incubate the calvariae for up to 14 d with medium changes every 3 d. 

3.3. Fixation 

1 . Fix the calvariae in 10% buffered formalin following termination of the cultures. 

2. Place the fixed calvariae between two sponge mats in a Tissue-Tek processing 
cassette and immerse overnight in decalcifying solution. 

3. The next day, process the demineralized calvaria overnight in a Shandon 
Hypercenter XP tissue-processing unit using the following protocol: 

70% Ethanol, 45 min. 
95% Ethanol, 45 min. 
95% Ethanol, 45 min. 
100% Ethanol, 45 min. 
100% Ethanol, 45 min. 
100% Ethanol, 45 min. 
100% Ethanol, 45 min. 
Xylene, 45 min. 
Xylene, 45 min. 
Paraffin, 45 min. 
Paraffin, 45 min. 

4. Embed the processed calvariae in paraffin blocks using a Tissue-Tek embedding 
console with paraffin wax. 

5. Orientate each calvaria with the midline suture down and each calvaria placed 
perpendicular to the plane of dissection and parallel to each other so that all 
calvariae from the same group are sectioned in the same orientation (Fig. 7; see 
also Note 4). 

3.4. Sectioning 

1. Trim the blocks to a depth of 800 urn to get past the remains of the midline 

14/Garrett/183-198/F1 188 1 2/26/03, 10:46 AM 

I ■ I I I 

In Vitro Bone Formation Assay 



Histological Sections cut 

4 Calvaria 
per block 



Fig. 7. Overview of the procedure of the histological sectioning of the half neonatal 
murine calvaria. Sections are cut parallel to the midline suture and include a cut 
through the coronal suture. Sections show both the front and parietal bones from the 
calvaria with the coronal suture area visible between the two bones. 

2. Cut 5-[iM thick sections. 

3. Cut further 5-u.m sections are taken at 400-u,m and 800-p.m depths. 

4. Mount the sections onto treated glass microscope slides. 

5. Air-dry in a 40°C oven overnight in preparation for staining. 

3.5. Staining 

Deparaffinize and stain the sections in the tissue processor using the follow- 
ing protocol: 

1. Three changes in xylene, 2 min. 

2. Three changes in 100% ethanol 1 min. 

3. 95% Ethanol, 30 sec. 

4. 80% Ethanol, 30 sec. 

5. Deionized water, 3 min. 

6. Hematoxylin solution, 1 min. 

7. Rinse in deionized water until section is clear. 

8. 0.5% Ammonia water, 10 sec. 

9. Rinse in deionized water, 3 min. 
10. 80% Ethanol, 30 sec. 



2/26/03, 10:46 AM 

I ■ I I I 


190 Garrett 

11. 95%Ethanol, 30 sec. 

12. Eosin solution, 3 min. 

13. Four changes in 95% ethanol, dips. 

14. Three changes in 100% ethanol, dips. 

15. Coverslip with Cytoseal XYL. 

3.6. Histomorphometry 

1 . The eosin Y differentially stains the original bone darker whereas the new bone 
matrix that is formed appears as a lighter color (Fig. 8A). The hematoxylin stains 
bone lining cells, active osteoblasts and osteocytes. 

2. Using the imaging processing equipment (see Subheading 2.6.), capture several 
images along the bone surface of the bone sections in each group. 

3. Outline the original bone using a pen-based drawing tablet attached to the 
computer (Fig. 8B; shaded area represents original bone present at the start 
of culture). 

4. Measure the amount of old bone and total bone using the color recognition option 
of the Image Pro software. 

5. Calculate the amount of new bone formed by subtracting the original bone from 
the total bone (see Note 5). 

6. Quantitate the bone lining cells by counting the number of cells that lie on each 
side of the calvaria. 

7. Measure the length of the bone represented by the image and calculate the num- 
ber of cells/mm of bone by dividing the cell number by the section length. 

8. Use a pen-based drawing tablet to measure the distance between the two outer 
edges of the suture. Suture width appears to increase with treatment anabolic 
agents that stimulate bone formation (Fig. 13; see Note 6). 

4. Notes 

1. Effects of growth factors in assay: We have investigated the anabolic effect of 
several growth factors in this assay (Fig. 9-11) . We find that BMP2, insulin, 
IGF-1, TGF-|3, fibroblast growth factor-acidic (FGFa), fibroblast growth factor- 
basic (FGFb), leukemia inhibitory factor (LIF) (Figs. 9 and 10) and endothelin-1 
(not shown here) all stimulate new bone formation over a 4-d period. We find 
that fibroblast growth factors increase bone resorption as well as increasing bone 
formation. Factors such as PDGFab, EGF, hMSP, bECGF, nerve growth factor 
(NGF), MIF, VEGF, and HGF (Figs. 10 and 11) did not stimulate bone formation 
in the assay. Prostaglandin E 2 (PGE 2 ) stimulates formation of new matrix but the 
major effect is to stimulate bone resorption. 

2. Time course and duration of assay: Examples of time course experiments are 
shown in Figs. 12 and 13. The calvariae were treated with medium alone or with 
medium plus 1 \iM simvastatin and terminated at various times after the start of 
the assay. Images of the calvarial bones were obtained from both the coronal 
suture area (Fig. 13) and from bones distant from the coronal suture (Fig. 12). 
Figure 13 illustrates the effect of either media alone (control) or simvastatin on 

14/Garrett/183-198/F1 190 1 2/26/03, 10:46 AM 

I ■ I I I 

In Vitro Bone Formation Assay 


Original Bone 

New Bone 

new ffljut." ^ < **VM1k 

Day 4 ?*?'* ?«^ ; ' : SlS^f? ; #^J': 

Original Bone ' 
Original Bone 


New Bone 

W -> 

•"••"'■i *W\?J^Bl?- < ; > 





Day 4 

Afew Bone 



Day 7 

Fig. 8. Neonatal murine calvaria (A) treated with either vehicle (Control) or 1 \jlM 
simvastatin for 4 d and 1 u,Msimvastatin for 7 d. Note new bone growth and increased 
number of active osteoblasts in simvastatin treated bones. (B) The same sections out- 
lined with hatched areas indicating the original bone from the start of the assay. This 
can be measured along with the total area of bone and the new bone area calculated. 

bone at the coronal suture of the calvaria. The suture area, which is evident in 
these histological sections as two layers, increases in size progressively from d 2 
to d 7 in control cultures. By d 7 the suture area appears to be relatively inactive. 



2/26/03, 10:46 AM 

I ■ I I I 






20ug/ml | 



50ng/mi 'jp^f 

■ X * \ 


TGFb r '1^;^"'! 
50ng/ml ,-:^\,'j- 


Fig. 9. Effects of various different growth factors on neonatal murine calvaria in- 
cluding 50 ng/mL of BMP2, 20 [xg/mL of insulin, 50 ng/mL of IGF-1, 50 ng/mL of 
TGF(3, and 50 ng/mL of acidic FGF. All stimulated bone formation in these cultured 
neonatal calvaria. 

In cultures treated with 1 \iM simvastatin, marked stimulatory effects on cellular- 
ity can be observed as early as d 1. Thereafter, there is a progressive increase the 
thickness and cell content of the coronal suture. Figure 12 compares the time 
course of cultures exposed to media alone with those exposed to simvastatin in 
the parietal bones distant from the coronal suture. The calvariae at this location 
are initially very thin pieces of bone that slowly increase in thickness over 7 d. 
You can see a small amount of new bone present in the control cultures by d 7. 



2/26/03, 10:46 AM 

I ■ I I I 

In Vitro Bone Formation Assay 



50ng/ml J 

50ng/ml ^.j^s.-. 

w> -JtSTJH^vwtf«ei^^ >m 

.Z^J" ! 



W*** -r *^'^" 



. —?&l.^j&Zi,-£cZZ 


" - 1 '- ? **^ 






Fig. 10. Effects of various different growth factors on neonatal murine calvaria 
including 50 ng/mL of basic FGF, 50ng/mL of LIF, 50 ng/mL of PDGFab, 50 ng/mL 
of EGF, 50 ng/mL of hMSP, and 50 ng/mL of bECGF. Of these, only basic FGF and 
LIF were able to stimulate new bone formation in this assay. 

When these calvariae are treated with an anabolic agent such as 1 \iM simvastatin, 
marked changes can be observed over the 7-d period. Noticeable effects can be 
observed as early as d 2 (48 h after the initial exposure to the compounds). This 
effect becomes more pronounced by d 7, when there are substantial increases in 
the number of cells on the bone with a marked increase in new bone matrix. 
Measurements of new bone matrix formation, suture width, and numbers of cells 
along the bone surface in these calvarial bones for the assay described in the 
preceding are summarized in Fig. 14. Simvastatin treatment caused a significant 



2/26/03, 10:46 AM 

I ■ I I I 











Fig. 11. Effects of various different growth factors on neonatal murine calvaria 
including 100 ng/mL of NGF, 100 ng/mL of MIF, 100 ng/mL of VEGF, 100 ng/mL of 
HGF, and 0.4 \iM PGE 2 . None of these stimulated bone formation. PGE 2 stimulated 
bone resorption in the cultured calvaria. 

increase in the amount of new bone formed by d 4, which was further increased by 
d 7. There was also a progressive increase in the suture width as well as the number 
of cells lining the bone, with the most significant changes seen at d 4 and 7. 

3. Storage of Harris hematoxylin: Do not keep the solution for more than 1 wk and 
filter each day before use. 

4. Mounting and sectioning the calvariae: Orientation of the calvariae in the paraf- 
fin block is very important for consistent results. If several calvariae from a single 
experimental group are going to be placed into one block, they must be oriented 
so that they are all perpendicular to the plane of sectioning and parallel to each 
other. They should also be set at the same depth in the block. An alternative 
procedure is place only one calvaria per block. This is initially easier, but it cre- 
ates much more work later during sectioning. 

5. Measurement of alkaline phosphatase: Cells within the calvariae express alkaline 
phosphatase, and this can be measured in conditioned medium using standard 
assays to give a semiquantitative assessment of osteoblast activity. 


2/26/03, 10:47 AM 

I ■ I I I 


In Vitro Bone Formation Assay 









Fig. 12. Histological sections away from the coronal suture of neonatal murine 
calvaria cultured with or without 1 \iM simvastatin for different times from d (day of 
dissection) to d 7. Simvastatin increased new bone matrix over the 7 d of culture. 

6. Analysis of suture width: The suture appears to be important target for stimula- 
tion of new bone in this assay. It may contain a number of cells that are target 
cells for either growth factors or small molecular weight anabolic compounds. 


1. Reynolds, J. J. (1967) The synthesis of collagen by chick bone rudiments in vitro. 
Exp. Cell Res. 47,42-48. 

2. Reynolds, J. J. (1968) Inhibition by calcitonin of bone resorption induced in vitro 
by vitamin A. Proc. R. Soc. Lond. B. Biol. Sci. 170, 61-69. 

3. Reynolds, J. J. (1966) The effect of ascorbic acid on the growth of chick bone 
rudiments in chemically defined medium. Exp. Cell Res. 42, 178-188. 

4. Reynolds, J. J. and Dingle, J. T. (1968) Time course of action of calcitonin on 
resorbing mouse bones in vitro. Nature. 218, 1 178-1 179. 

5. Minkin C, Reynolds J. J., and Copp, D. H. (1971) Inhibitory effect of salmon and 
other calcitonins on calcium release from mouse bone in vitro. Can. J. Physiol. 
Pharmacol. 49, 263-267. 



2/26/03, 10:47 AM 

I ■ I I I 










Fig. 13. Histological sections of the coronal suture of neonatal murine calvaria cul- 
tured with or without 1 \iM simvastatin for different times from d (day of dissection) 
to d 7. Simvastatin increased the width of the suture over the 7 d of culture. 

6. Reynolds, J. J. (1966) The effect of hydrocortisone on the growth of chick bone 
rudiments in chemically defined medium. Exp. Cell Res. 41, 174-189. 

7. Reynolds, J. J, Minkin, C, Morgan, D. B., Spycher, D., and Fleisch, H. (1972) 
The effect of two diphosphonates on the resorption of mouse calvaria in vitro. 
Calcif. Tissue Res. 10, 302-313. 

8. Gowen, M., Wood, D. D., Ihrie, E.J., McGuire, M. K., and Russell, R. G. (1983) 
An interleukin 1 like factor stimulates bone resorption in vitro. Nature 306, 

9. Bertolini, D. R., Nedwin, G. E., Bringman, T. S., Smith, D. D., and Mundy, G. R. 
(1986) Stimulation of bone resorption and inhibition of bone formation in vitro by 
human tumor necrosis factors. Nature 319, 516-518. 

10. Lorenzo, J. A., Sousa, S. L., and Leahy, C. L. (1990) Leukemia inhibitory factor 
(LIF) inhibits basal bone resorption in fetal rat long bone cultures. Cytokine 2, 

11. Reynolds, J. J. and Dingle, J. T. (1968) The induction and inhibition of bone 
resorption in vitro. Calcif. Tissue Res. (Suppl) 50-50a. 



2/26/03, 10:47 AM 

I ■ I I I 

In Vitro Bone Formation Assay 




0.20 - 

0.15 - 

0.10 - 

0.05 - 


200 - 

* 160 - 

| 160 - 

•5 140 - 

£ 120 - 


80 - 

60 - 

40 - 




12 4 7 

Days of incubation 

Fig. 14. Histomorphometric measurements of new bone area, suture width and num- 
bers of cells in neonatal murine calvaria cultured with or without 1 u,M simvastatin for 
different times from d (day of dissection) to d 7. Simvastatin significantly increased 
new bone matrix, numbers of cells/mm bone, as well as suture width, over the 7 d of 



Reynolds, J. J. and Dingle, J. T. (1970) A sensitive in vitro method for studying 
the induction and inhibition of bone resorption. Calcif. Tissue Res. 4, 339-349. 
Traianedes, K., Dallas, M. R., Garrett, I. R., Mundy, G. R., and Bonewald, L. F. 
(1998) 5 -Lipoxygenase metabolites inhibit bone formation in vitro. Endocrinol- 
ogy 139, 3178-3184. 



2/26/03, 10:47 AM 


I ■ I I I 

198 Garrett 

14. Mundy, G, Garrett, R, Harris S, et al. (1999) Stimulation of bone formation in 
vitro and in rodents by statins. Science 286, 1946-1949. 

15. Garrett, I. R., Esparza, J., Chen, D., et al. (2000) Statins mediate their effects on 
osteoblasts by inhibition of HMG-CoA reductase and ultimately BMP-2. /. Bone 
Miner. Res. 15S, 225. 

16. Garrett, I. R., Chen, D., Zhao, M., et al. (2001) Statins mediate bone formation by 
enhancing BMP2 expression. Bone 28, 75. 

17. Yin, J. L., Grubbs, B. G., Cui, Y., et al. (2000) Endothelin A receptor blockade 
inhibits osteoblastic metastasis. /. Bone Miner. Res. 15S, 201. 

18. Garrett, I. R., Gutierrez, G., Chen, D., et al. (2000) Specific inhibitors of the chy- 
motryptic component of the proteasome are potent anabolic agents in vivo. /. 
Bone Miner. Res. 15S, 197. 

19. Chen, D., Garrett, I.R., Qiao, M., et al. (2001) Proteasome inhibitors stimulate 
osteoblast differentiation and bone formation by inhibiting Gli3 degradation and 
enhancing BMP-2 expression. Bone 28, 74. 

14/Garrett/183-198/F1 198 1 2/26/03, 10:47 AM 

I ■ I I I 



In Situ Hybridization and In Situ 

Reverse Transcription Polymerase Chain Reaction 

in Human Bone Sections 

Andrew P. Mee and Judith A. Hoyland 

1 . Introduction 

Gene expression in bone can be assessed by several techniques such as 
reverse transcription polymerase chain reaction (RT-PCR), differential dis- 
play PCR, subtractive hybridization, and microchip arrays. The problem with 
all of these techniques is that they do not allow cellular localization of the 
genes under investigation. In situ hybridization (ISH) circumvents this prob- 
lem. It is possible to localize precisely sites of gene expression using ISH in 
intact cells and tissue sections, allowing one to build up a more complete pic- 
ture of the processes occurring in the disease under investigation. 

Although ISH is a very powerful technique for localizing expression of most 
targets, low levels of DNA or RNA (fewer than 20-40 copies per cell) cannot 
easily be detected (reviewed in ref. 1). In view of this, PCR has been devel- 
oped to increase the sensitivity of ISH. For DNA targets, a PCR step is per- 
formed to increase the abundance of target, prior to its detection by ISH. This 
technique is termed in situ PCR (IS-PCR). For RNA targets, a reverse tran- 
scription step is performed prior to PCR-based amplification of the target mol- 
ecule with final detection by ISH (IS-RT-PCR). The fact that amplification 
during IS-PCR and IS-RT-PCR occurs within the cells of interest reduces the 
possible problem of contamination associated with conventional PCR. IS-PCR 
has been used primarily to detect viral DNA (2-5) and single copy genes (6) 
within cells, whereas IS-RT-PCR has been used to detect expression of a vari- 
ety of genes in bone and other tissues (7-19). Although ISH and IS-RT-PCR 
are useful techniques, they are technically difficult and time-consuming. Here 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


15/Mee/199-224/F1 201 1 2/26/03, 10:47 AM 

I ■ I I I 


202 Mee and Hoyland 

we review the application of ISH and IS-RT-PCR to the detection of gene 
expression in bone. 

2. Materials 

2.1. Equipment 

1. ThermoHybaid Omnislide system (Fig. 1): For thermal cycling in IS-RT-PCR. 

2. ThermoHybaid Easiseals: To seal reactions during ISH-RT-PCR. 

3. Tissue processor: For tissue processing. 

4. Microtome: For tissue sectioning. 

5. Small artist's paintbrush: For transferring sections to slides. 

6. Microscope slides and coverslips. 

7. Metal and plastic slide racks and glass slide troughs. 

8. Lightproof slide box for autoradiography. 

9. Developing tanks for autoradiography. 
10. Plastic forceps for autoradiography. 

2.2. Tissue Fixative 

1. 10 mL of 37% formaldehyde in 90 mL of phosphate-buffered saline (PBS). 

2.3. Decalcification Solution 

1. 20% (w/v) Ethylenediaminetetraacetic (EDTA) acid in water. Adjust to pH 7.2 
with Na OH. 

2.4. DEPC Water 

1. Add 1 mL of diethylpyrocarbonate (DEPC) (Caution: This compound is carci- 
nogenic.) to 999 mL of distilled water. Incubate for 4 h at room temperature and 
autoclave to inactivate DEPC. 

2.5. General Reagents 

1. 0.5 M EDTA, pH 8.0: Add 93.05 g of EDTA to 300 mL of DEPC-treated water 
and adjust pH to 8.0 with 10 M NaOH. Make up to 500 mL with DEPC-treated 
water. Autoclave to sterilize. 

2. 1 M DTT: Add 3.09 g of dithiothreitol (DTT) to 20 mL of DEPC-treated water, 
sterile filter and store in 1-mL aliquots at -20°C until use. 

3. 1 M Tris-HCl, pH 8.0: Add 121 g of Tris base to 800 mL of DEPC-treated water; 
adjust to pH 8.0 with concentrated HC1. Make up to 1 L with DEPC-treated water 
and autoclave to sterilize. 

4. 20x Saline sodium citrate (SSC) buffer: Add 175.3 g of NaCl and 88.2 g of so- 
dium citrate-H 2 to 800 mL of DEPC-treated water. Adjust the pH to 7.0 with 1 M 
HC1, and make up to 1 L with DEPC-treated water. Autoclave to sterilize. 

5. STE buffer: 0.1 M of NaCl, 10 mM Tris-HCl, pH 8.0, 1 mM ETDA, pH 8.0. 

6. NTE buffer: 0.5 M NaCl, 10 mM Tris-HCl, pH 8.0, and 1 mM EDTA, pH 8.0. 

15/Mee/199-224/F1 202 I 2/26/03, 10:47 AM 

I ■ I I I 

In Situ Hybridization and In Situ RT-PCR in Bone 



Fig. 1. Omnislide™ system and Easiseals. (A) The Omnislide is a flat-block system 
that holds 20 slides in a plastic slide rack. Hence, all slides can be manipulated at once, 
thus making handling much easier. Further, many of the incubations can be carried out 
by pipetting solutions directly onto the slides, without the need for coverslips. The 
system includes wash sleeves, an ambient sleeve rack, and a heated module for per- 
forming stringency washes within the wash sleeves. (B) Easiseals shown in a variety 
of sizes. The Easiseal consists of a plastic frame that is adherent on both sides, and a 
clear plastic cover that is laid over the frame on a slide to form an airtight seal. 

7. 3 M Sodium acetate: Add 401.8 g of NaAc. 3H 2 to 800 mL of DEPC-treated 
water. Adjust pH to 5.2 with glacial acetic acid and make up to 1 L with DEPC- 
treated water. Autoclave to sterilize. 

8. 50x Denhardt's solution: Dissolve 5 g of Ficoll, 5 g of polyvinylpyrrolidine, and 
5 g of bovine serum albumin into 500 mL of DEPC-treated water. Filter sterilize 
and store frozen in aliquots of 1-5 mL. 

9. 50x Modified Denhardt's solution: Dissolve 5 g of Ficoll, 5 g of 
polyvinylpyrrolidine, and 25 g of bovine serum albumin into 500 mL of DEPC- 
treated water. Filter sterilize and store frozen in aliquots of 1-5 mL. 

2.6. Materials for Tissue Embedding, Sectioning, 
and Slide Preparation 

1. Ethanols: 50%, 70%, 90%, 99%, and 100%. 

2. Xylene. 

3. APES solution: y-Aminopropyltriethoxysilane (APES) diluted to 3% in acetone. 

4. Acetone. 

5. Industrial methylated spirits (IMS). 

6. 1MHC1. 

7. 5% Dimethyldichlorosilane in xylene: Add 5 mL of dimethyldichlorosilane to 95 mL 
of xylene. 

2.7 Preparation of Radioactive cDNA Probes 

1 . Amersham Megaprime™ kit. 

2. [ 35 S] dCTP. 

3. EDTA stop buffer: 0.2 M EDTA, pH 8.0. 



2/26/03, 10:47 AM 

I ■ I I I 


204 Mee and Hoyland 

2.8 Preparation of Digoxigenin(DIG)-Labeled cDNA Probes 

1. DIG High Prime DNA labeling and detection starter kit II (Roche Molecular 

2. EDTA stop buffer: 0.2 M EDTA, pH 8.0. 

2.9. Preparation of Radioactive Riboprobes 

1. Riboprobe labeling kit (Promega). 

2. 10. Preparation of DIG-labeled Riboprobes 

1. DIG RNA labeling Kit (SP6/T7) (Roche). 

2.11. Quantification and detection of DIG-Labeied Probes 

DIG quantification test strips (Roche). 

DIG control test strips (Roche). 

Dilution buffer: 50 ixg/mL of herring sperm DNA in 10 mM Tris-HCl, pH 8.0. 

Maleic acid Buffer: 0.1 M Maleic acid, 0.15 M NaCl, pH 7.5. 

Blocking buffer: Dilute blocking solution 1:10 with maleic acid buffer. 

Detection Buffer: 0. 1 M Tris-HCl, 0. 1 M NaCl, 50 mM MgCl 2 , pH 9.5. 

Anti-DIG AP antibody solution: Dilute antibody stock 1 :2000 in maleic acid buffer. 

Substrate solution: Add 40 \iL nitroblue tetrazolium/5-bromo-4-chloro-3- 

indolylphosphate (NBT/BCIP) stock solution to 2 mL of detection buffer. 

2. 12. Sephadex Mini-Spin Columns 

To separate unincorporated nucleotides from cDNA probes: 

1. Plug a 1-mL syringe with glass wool. 

2. Fill the plugged syringe with Sephadex G-50 suspended in STE buffer. 

3. Place the syringe in a 15-mL centrifuge tube and centrifuge at 1500 rpm for 
1 min to pack the column. 

4. Pour off the STE from the centrifuge tube. 

2. 13. cDNA Probe Hybridization Buffer 

For 1 mL combine together in an Eppendorf tube: 

a. 0. 1 g of dextran sulfate 

b. 100 u,L of lOx modified Denhardt's solution. 

c. 200 uL of 3 M NaCl. 

d. 20 \iL of 10 mg/mL salmon sperm DNA. 

e. 10 uL of 1 M Tris-HCl, pH 7.4. 

f. 1 ixL of 0.5 M EDTA. 

g. lOnLoflMDTT. 

h. 500 \iL of deionized formamide. 
i. RNase-free water to 1 mL. 

Incubate at 37°C until the reagents have dissolved. 

15/Mee/199-224/F1 204 I 2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 205 

2. 14. Riboprobe Hybridization Buffer 

1. For 1 mL, combine together in an Eppendorf tube: 

a. 500 \xL of deionized formamide. 

b. 60 \ih of 5 M NaCl. 

c. 20 uE of 1 M Tris, pH 8.0. 

d. 10 \ih of 0.5 M EDTA, pH 8.0. 
c. 10 uE of 1 M DTT. 

2. Warm to 60°C and add: 

a. 100 mg of dextran sulfate. 

b. 20 \iL of 50x Denhardt's solution. 

c. 100 \iL of 10 mg/mL tRNA (E. coli). 

d. 230 \iL of DEPC-treated water. 

2. 15. Alkaline Hydrolysis Buffer 

1. For 1 mL combine 500 [iL of 0.4MNaHCO 3 and 500 \iL of 0.6MNa 2 CO 3 , pH 10.0. 

2. 16. Autoradiography and Counterstaining 

1. Photographic emulsion: Ilford K5 emulsion. 

2. Developer: Ilford D-19 developer. 

3. Fixative: Ilford Hypam. 

4. Filtered Harris' hematoxylin and 2% eosin. 

5. Loctite UV adhesive (glass bond). 

2.17. Indirect IS-RT-PCR 

1. Random hexamer solution: Add 15 \ig of random hexamers to 100 [iL of DEPC- 
treated water. 

2. AMV reverse transcription system (Promega). 

3. PCR amplification mix: 5 U of Tctq DNA polymerase; 5 uE of lOx Taq DNA 
polymerase buffer; lOpmol of each primer; 0.2 mM each of dATP, dCTP, dGTP, 
and dTTP, distilled water to 50 |xL. 

2.18. Direct IS-RT-PCR 

This can use the Tth enzyme (20), which possesses both RT and DNA poly- 
merase activity: 

1 . Reaction mix: 10 jjE of 5x reaction buffer; 5 uL of 25 mM Mn(OAc) 2 ; 6 [ih of 10 mM 
dNTP mix; 0.65 \iL of 10 mM DIG dUTP; 0.5 pmol of each primer; 10 U of rTth 
enzyme; add sterile distilled water to 50 \iL. 

3. Methods 

3. 1. Fixation, Decalcification, and Embedding 

In order to obtain satisfactory paraffin sections of bone, it is necessary to 
remove the mineral and then soften the tissue. This can be carried out with 

15/Mee/199-224/F1 205 1 2/26/03, 10:47 AM 

I ■ I I I 


206 Mee and Hoyland 

acids or chelating agents. The chelating agent EDTA in the form of its diso- 
dium salt has minimal effect on mRNA retention, whereas acids greatly reduce 
mRNA (21) and should not be used. 

1. Transport the sample to the laboratory within 30 min of retrieval and cut to a 
manageable size if necessary (see Note 1). 

2. Immerse the sample in fixative in a universal container and incubate for 3-48 h at 
4°C depending on the sample size (see Note 2). 

3. Wash the sample in distilled water and transfer to EDTA for 10-14 d, with fresh 
changes of EDTA every 3 d (see Note 3). 

4. Dehydrate the tissue, through a series of ethanols and xylene and embed in paraf- 
fin wax (see Note 4). 

3.2. Tissue Sectioning 

We usually cut bone sections of 7 [xm thickness. If samples are fragile, we 
cut sections onto sellotape as described below, as this increases adherence of 
the section to the slide and is particularly suitable for calcified tissues. 

1 . Place a strip of adhesive tape onto the surface of the block and press down firmly, 
leaving a sufficient length of tape free in front of the block. 

2. Raise this end above the level of the knife edge and cut a section in the usual 
manner. The tissue section will adhere to the tape. 

3. Continue until sufficient sections are obtained. 

4. Cut round the section and press down firmly onto a clean subbed slides using water. 

5. Dry on hot plate (60°C) for 1-7 d. 

6. Remove the tape by placing the slide in acetone for 1 min. 

7. Remove adhesive by placing in chloroform for 20-30 min. 

3.3. Subbing Slides 

Slides are subbed to increase adherence of cells and tissue sections to the 
slides (22). 

1. Place slides in a metal rack and wrap in aluminum foil. 

2. Bake at 250°C for 3 h. 

3. Wearing gloves, unwrap the slides and allow them to cool slightly. Rinse in IMS 
and air-dry. 

4. Immerse slides in APES solution for 5 min. 

5. Immerse slides in acetone. 

6. Rinse slides in DEPC-treated water. 

7. Place in oven at 37°C overnight. 

8. Store in dust-free conditions until use. 

3.4. Mounting Sections on Slides 

1 . Float the sections onto DEPC-treated water at 50°C. Any wrinkles in the sections 
can be removed by gently pressing on the section with a small paintbrush. 

15/Mee/199-224/F1 206 1 2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 207 

2. Introduce a subbed microscope slide (see Subheading 3.3.) into the water under- 
neath the floating section, at an angle of approx 45° and carefully lift the slide to 
the surface of the water so that the section adheres to the slide. 

3. Place the slides on a hot plate (60°C) to dry overnight (for ISH; see Note 5) or up 
to 5 d (for IS-RT-PCR). 

3.5. Silanation ofCoverslips 

Silanation of coverslips minimizes the amount of hybridization buffer 
needed during ISH. Such coverslips can be purchased commercially (Sigma 
Hybrislips), or prepared using the protocol below. 

1. Place coverslips in metal coverslip racks. 

2. In a fume cupboard wash the coverslips in IMS for 5 min. 

3. Allow to air-dry. 

4. Wash the coverslips in 1 M HC1 for 20 min. 

5. Rinse in DEPC-treated water. 

6. Dry in a 60°C oven. 

7. In a fume cupboard, immerse the coverslips in 5% dimethyldichlorosilane in 
xylene for 5 min. 

8. Allow to air-dry in fume cupboard. 

9. Bake in 200°C oven overnight. 

10. Rinse in DEPC-treated water. 

1 1 . Dry in a 60°C oven. 

3.6. Choice of Probes for ISH 

Many different probes and labeling systems can be used for ISH, each with 
different advantages and drawbacks (see Note 6). The following protocols give 
details of making radioactively labeled and DIG-labeled cDNA probes and 

3.7. Random-Primed Radioactive cDNA Probes 

The labeling reaction described below (50 ng of probe) produces sufficient 
labeled probe to hybridize to 10 tissue sections. 

1. Combine 50 ng of probe and 10 u,L of a random hexamers mix in a 1.5 mL 
Eppendorf tube. Make up the reaction volume to 52 u,L with sterile distilled water 
and pierce a hole in the lid of each Eppendorf tube. 

2. Incubate at 100°C for 5 min to denature the double-stranded DNA probe. Allow 
the mixture to cool to room temperature. 

3. Add the following reagents from the Megaprime kit to the tube: 

a. 10 fxL of reaction buffer. 

b. 8 u,L of each unlabeled dNTP. 

c. 10 uE of [ 35 S]-dCTP (50 M-Ci). 

d. 4 u,L of Klenow fragment DNA polymerase. 

4. Incubate at 37°C for 1 h and terminate the reaction by adding 5 uE of EDTA stop 
buffer to the tube. 

15/Mee/199-224/F1 207 I 2/26/03, 10:47 AM 

I ■ I I I 


208 Mee and Hoyland 

5. Transfer the reaction to a Sephadex G-50 mini-spin column (see Subheading 
2.12) and remove the unincorporated nucleotides by centrifuging at 10,000g in a 
microcentrifuge for 1 min. 

6. Use the labeled probe immediately or freeze at -20°C for 24-48 h until use. 

3.8. Nonradioactive (DIG) Random-Prime Labeling of cDNA 

1 . To an Eppendorf tube containing 1 jxg of cDNA probe, add sterile water to a final 
volume of 16 \iL. 

2. Pierce a hole in the lid of the Eppendorf tube and incubate at 100°C for 5 min to 
denature the DNA. 

3. Place on ice and immediately add 4 jxL of DIG high prime mix. Vortex to mix 
and centrifuge briefly. 

4. Incubate overnight at 37°C. 

5. Stop the labeling reaction by adding 2 [iL of EDTA stop buffer. 

6. Use immediately or store at -20°C for up to 12 mo. 

3.9. Quantitating DIG Probe Concentration 

It is important to quantify the amount of labeled probe that has been gener- 
ated when using DIG-labeled probes (see Note 7). Too much probe causes 
high background, whereas too little leads to a weak signal. The amount of probe 
can be estimated using test strips as described here: 

1. Add 1 (iL of labeled probe to 19 \xL of DEPC-treated water to give a stock solu- 
tion of approx 1 ng/fxL. 

2. Make serial dilutions of this solution at 1:3; 1:10; 1:30; 1:100; and 1:300 using 
dilution buffer: 

3. Dot 1 \iL of each dilution on the DIG quantification test strip. 

4. Place the test strips and the DIG control strip in blocking buffer for 2 min. 

5. Transfer the strips to antibody solution for 3 min. 

6. Transfer to blocking buffer for 1 min. 

7. Transfer to maleic acid buffer for 1 min. 

8. Transfer to detection buffer for 1 min. 

9. Transfer to substrate and incubate in the dark, for 5-30 min, checking for color 
development to monitor progress of the reaction. 

10. Estimate the amount of DIG-labeled probe present by comparing with the DIG 
control test strip. 

3. 10. Radioactive Labeling of Riboprobes 

1. Add the following together in an Eppendorf tube: 

a. 1 u,g of linearized plasmid template. 

b. 4 |iL of 5x transcription buffer. 

c. 2 \xL of 100 mM DTT. 

d. 1 \iL each of rATP, rCTP, rGTP. 

15/Mee/199-224/F1 208 1 2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 209 

e. 2.4 uX of 100 [iM rUTP. 

f. 1 |xL of RNasin. 

g. 5 uE (50 ,uCi) of [ 35 S]UTP. 

h. 1 |xL SP6, T3, or T7 polymerase. 

2. Incubate for 2 h at 37°C and stop the reaction by adding 2 [xL of EDTA stop 

3. Remove unincorporated nucleotides and plasmid DNA by adding 4 |xL of DNase 
and 1.5 [xL of RNasin to the tube and incubate at 37°C for 15 min. 

4. Add 50 |xL of 7.5 M ammonium acetate and 375 u,L of cold 100% ethanol to the 
tube. Place at -20°C overnight or at -70°C for 2-4 h to precipitate the RNA. 

5. Centrifuge at 10,000g for 20 min. Aspirate the ethanol with care and dispose of 

6. Wash the RNA pellet with 500 \iL of 70% ethanol. 

7. Centrifuge for 10 min at 10,000g. Remove ethanol and allow the pellet to dry. 

8. Suspend the probe in 19 uE of 100 mM DTT and 1 uE of RNasin. 

9. For riboprobes of > 500 basepairs (bp), proceed to limited alkaline hydrolysis 
(see Subheading 3.12.) before use (see Note 8). 

10. Add 1 u,L of probe to 3 mL of scintillation fluid and read counts on the scintilla- 
tion counter. For satisfactory results, the probe should have >600,000 counts/uE. 

3.11. Nonradioactive (DIG) Labeling of Riboprobes 

1. To 1 |xg of linearized DNA add 2jxL of lOx buffer, 2 [xL of NTP mix, 1 (xL of 
RNase inhibitor, and 2 |xL of the appropriate RNA polymerase (T7, T3, or SP6). 
Make the total volume up to 20 |xL with DEPC-treated water. 

2. Place the reaction in a microcentrifuge and pulse spin for 10 sec to ensure reagents 
are mixed. 

3. Incubate at 37°C for 2 h. 

4. Add 2 jxL of DNase and 1 jxL of RNasin and incubate at 37°C for another 15 min. 

5. Stop the reaction by adding 2 |xL of EDTA stop buffer. 

6. If probes are longer than 500 bp, reduce to 500 bp by limited alkaline hydrolysis 
(see Subheading 3.12.). 

7. Add 500 jxL of cold ethanol to the tube, place at -20 C C and leave overnight. 

8. Spin the tube at 10,000g for 20 min in a microcentrifuge. 

9. Remove the supernatant, taking care not to disturb the RNA pellet. 

10. Leave to air-dry (approx 30-60 min). 

11. Resuspend pellet in 19 |xL of DEPC-treated water and 1 uE of RNasin. 

12. Store at -20°C for up to 12 mo until use. 

3. 12. Alkaline Hydrolysis of RNA Probes 

1. Resuspend the probe in 160 [xL of DEPC-treated water. 

2. Add 40 \xh of hydrolysis buffer and incubate at 60°C for a time defined by the 

t = L - Lf I KLgLf 

15/Mee/199-224/F1 209 I 2/26/03, 10:47 AM 

I ■ I I I 


210 Mee and Hoyland 

where t = time in min, L = initial probe length in kb, L f = desired final probe 
length in kb and K = rate constant of hydrolysis = 0.11. 

3. Stop the reaction by adding 6.6 uE of 3 M sodium acetate and 1.3 uE of glacial 
acetic acid. 

4. Precipitate the reaction by adding 500 uE of ethanol and leave at -20°C for at 
least 2 h or overnight. 

5. Centrifuge at 10,000g for 15 min, aspirate the supernatant, and allow pellet to dry. 

6. Resuspend the probe in 19 ^L of 100 mMDTT (for 35 S-labeled probes) or DEPC- 
treated water (for DIG-labeled probes) and 1 jxL of RNasin. 

7. Store at -20°C for up to 4 wk ( 35 S-labeled probes) or 12 mo (DIG-labeled probes) 
until use. 

3. 13. Prehybridization ISH and IS-RT-PCR 

1. Dewax the sections by placing in xylene for 5 min. Repeat three times. 

2. Transfer the sections to 95% IMS for 5 min. Repeat four times. 

3. Wash the sections with DEPC-treated water. 

4. Place the sections in 0.2 M HC1 for 20 min at room temperature. 

5. Wash in 2x SSC or water for 3 min. Repeat once. 

6. Rinse in 0.05 M Tris HC1, pH 7.4 for 3 min. 

7. Cover the section with 10 [xg/mL of Proteinase K solution to cover the entire 

8. For ISH, incubate the slides at 37°C for 1 h. For IS-RT-PCR, this time is reduced 
to 20 min {see Note 9). 

9. Rinse slides in 0.2% glycine-PBS, for 3 min. Repeat once. 

10. Rinse slides in PBS. 

11. If required, prepare negative controls by incubating selected slides with 10 mg/ 
mL of crude RNase for 60 min {see Note 10). 

12. If required, postfix in 4% paraformaldehyde-PBS for 4 min at room temperature. 

13. Rinse slides in PBS. 

14. Immerse in freshly prepared 0.25% acetic anhydride, 0.1 M triethanolamine, pH 
8.0, for 10 min {see Note 11). 

15. Rinse in DEPC-treated water and then dehydrate by incubating in 90% IMS for 5 min. 

16. Allow to air-dry and proceed to ISH {see Note 12). 

3. 14. ISH with Radioactive cDNA Probes 

1. Make up 1 mL of hybridization buffer and combine with two labeling reactions. 
Incubate at 100°C for 5 min and cool by placing on ice. 

2. Apply approx 50 fxL of probe mix to each tissue section. 

3. Cover sections with treated coverslips and incubate overnight at 37 °C. 

3.15. ISH with DIG-Labeled cDNA Probes 

Follow the same protocol as described in Subheading 3.14., using 500 ng of 
probe per milliliter of hybridization buffer. The amount of probe should be 

15/Mee/199-224/F1 210 1 2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 21 1 

estimated by the results of the DIG quantification test strips (see Subhead- 
ing 3.9.). 

3. 16. ISH with Radioactive Riboprobes 

1. Add sufficient quantity of probe (as determined by scintillation counting [see 
Subheading 3.10]) to riboprobe hybridization buffer (see Subheading 2.14.). 

2. Heat probe to 60°C for 10 min and cool on ice. 

3. Apply 20-50 uL (approx 0.1 x 10 6 dpm) probe mix per section (total counts 
required in 500 jxL is 2.5 x 10 6 dpm). 

4. Place siliconized coverslip over section and hybridize overnight at 50 C C. 

3.17. Non-radioactive Riboprobe Hybridization 

Follow the same protocol as described in Subheading 3.16., using 500 ng of 
probe per millilter of hybridization buffer. The amount of probe should be esti- 
mated by the results of the DIG quantification test strips (see Subheading 3.9.). 
Typical results of radioactive and nonradioactive ISH are shown in Fig. 2. 

3.18. Post-hybridization Washes forcDNA Probes 

The same protocol applies for both radioactive and nonradioactive probes. 
The temperature of the incubation in step 5 depends on the length of the probe 
and the G/C content (see Note 13). 

1. Remove the coverslips from the slides by soaking in 4x SSC for 5 min. 

2. Incubate the slides in 0.5x SSC, 1 mM EDTA, and 10 mM DTT for 5 min. Repeat once. 

3. Transfer to 0.5x SSC, 1 mM EDTA for 5 min. Repeat once. 

4. Transfer to 50% formamide, 0.15 M NaCl, 5 mM Tris-HCl, pH 7.4; 0.5 mM 
EDTA, pH 8.0, and incubate for 10 min in a fume cupboard. 

5. Transfer to 0.5x SSC at the probe T m minus 10°C for 5 min. Repeat three times. 

6. Transfer to 0.5x SSC and incubate at room temperature for 5 min. 

7. Transfer to IMS and incubate at room temperature for 5 min. 

8. Remove from IMS and allow to air-dry. 

3. 19. Post-hybridization Washes for Radioactive Riboprobes 

1. Rinse slides in 2x SSC-1 mM DTT until coverslips come off. 

2. Transfer to fresh 2x SSC and leave at room temperature for 60 min. 

3. Transfer to wash buffer for 4 h at 50°C. 

4. Transfer to NTE buffer for 2 min. 

5. Transfer to fresh NTE buffer containing 20 ng/mL of RNase A and 100 U/mL of 
RNase Tl for 30 min at 37°C. 

6. Transfer to NTE buffer and incubate for 30 min at 37°C. 

7. Transfer to wash buffer and leave overnight at 50°C. 

8. Transfer to 2x SSC and leave at room temperature for 30 min. 

9. Transfer to 0.1 x SSC and leave at room temperature for 30 min. 

15/Mee/199-224/F1 211 I 2/26/03, 10:47 AM 

I ■ I I I 


Mee and Hoyland 


D r ^ tV -\ - * 

A m : ■ 

p * • 

Fig. 2. Comparison of radioactive and nonradioactive ISH. (A) Transforming 
growth factor-|3 (TGF-(3) mRNA demonstrated by autoradiography. (B) TGF-|3 mRNA 
demonstrated by digoxigenin — disclosed by NBT/BCIP (dark brown/blue stain). (C) 
Type I collagen mRNA demonstrated by autoradiography. (D) Type I collagen mRNA 
demonstrated by digoxigenin. 

10. Transfer to IMS and incubate at room temperature for 5 min. 

11. Remove from IMS and allow to air-dry. 

3.20. Post-hybridization Washes for Nonradioactive Riboprobes 

1. Immerse slides in 4x SSC until the coverslips come off. 

2. Transfer to 50% deionized formamide in lx SSC and incubate at 55 °C for 
20 min. Repeat for a total of three times. 

3. Transfer to lx SSC and incubate for 15 min at room temperature. Repeat once. 

3.21. Autoradiography and Counterstaining 

The following steps should be done in a dark room. Before beginning, make 
sure you have your slides to be dipped, along with blank clean slides, a dipping 
chamber, plastic slide racks, plastic forceps, a clean developing tank lined with 


2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 213 

paper roll, and a water bath held at 4(M15°C. Do not use metal objects — this 
will "stress" the emulsion and can produce false positive results. 

1. Fill a universal container with 10 mL of distilled water and incubate at 40-45°C 
for 30 min. 

2. Fill a universal container with approx 10 mL of K5 emulsion, using plastic for- 
ceps, and put the lid on. Place in a water bath at 40-45°C until it melts. 

3. When the emulsion has melted add an equal volume of warm distilled water and 
invert slowly to mix. 

4. Pour the diluted emulsion in a clean dipping chamber and dip a few blank slides 
to remove any air bubbles. 

5. Dip the slides in the emulsion and allow them to drain upright on the side of the 
lined developing tank. 

6. Put the slides upright onto plastic slide racks and allow to dry in a dark cupboard 
for 3-4 h. 

7. When the slides are dry put them into a lightproof box, label, and seal the edges 
with tape. 

8. Incubate slides at 4°C for 7-10 d if riboprobes are being used and for 10-14 d if 
cDNA probes are being used. 

9. When the incubation is complete, take the slides to a darkroom, remove from the 
lightproof box, and transfer to a glass trough and cover with D-19 developer. 
Incubate for 5 min at 24°C. 

10. Rinse slides in three changes of distilled water. 

11. Cover slides with fixative and incubate for 5 min. 

12. Wash once in distilled water. 

13. Rinse in running tap water for 5 min. 

14. Immerse in filtered Harris' Hematoxylin for 2 min. 

15. Rinse in running tap water for 5 min. 

16. Immerse in 2% eosin for 30 sec. 

17. Rinse in water and allow to air-dry. 

18. Add two drops of Loctite UV adhesive to each section and put on a coverslip. 

19. Expose the slides to UV light using a UV transilluminator for 2 min. 

3.22. Detection of DIG-Labeled cDNA Probes and Riboprobes 

1 . Place slides in maleic acid buffer for 5 min. 

2. Place slides in a wet box and cover section with 0.5% Boerhinger Blocking 
Reagent (BBR). Incubate for 30 min at room temperature. 

3. Tip off BBR from sections, and cover with a 1-500 dilution of anti-DIG (AP) in 
0.5% BBR. Incubate for 2 h at room temperature. 

4. Wash slides in maleic acid buffer. 

5. Place slides in a wet box and equilibrate in detection buffer. Remove detection buffer 
and then cover section with substrate solution. Incubate in the dark and monitor the 
progress of the reaction over 1-12 h by looking for an appearance of staining. 

6. Wash slides in distilled water and allow to air-dry. 

7. Mount in Loctite UV adhesive as described in Subheading 3.21., steps 18 and 19. 

15/Mee/199-224/F1 213 1 2/26/03, 10:47 AM 

I ■ I I I 

Mee and Hoyland 


Fig. 3. Comparison of indirect and direct IS-RT-PCR. Note that these figures show 
sections of human kidney stained for the presence of vitamin D receptor (VDR) 
mRNA — we have never used the direct method on bone sections as it is such an unre- 
liable technique. (A) Direct IS-RT-PCR, 20 cycles, positive staining in renal tubules. 
Note the poor morphology due to cycling for 20 times at 95°C. (B) Direct IS-RT-PCR, 
20 cycles, negative control — omission of DNA polymerase. Note absence of staining 
(nuclei counterstained with methyl green). (C) Direct IS-RT-PCR, 20 cycles, negative 
control — omission of reverse transcriptase. Note presence of intense staining in all 
nuclei due to nonspecific DNA repair. It is also of note that the cytoplasm in (C) is not 
stained, hence the technique has worked in (A) (where the cytoplasm is also stained). 
However, the presence of staining in the nuclei means that this technique is unreliable. 
(D) Indirect IS-RT-PCR, five cycles, positive staining in renal tubules. Note the stain- 
ing is less intense than in (A) (only five cycles compared with 20), but that the mor- 
phology is much better in (D). (E) Indirect IS-RT-PCR, five cycles, negative 
control — omission of DNA polymerase. Note absence of staining (nuclei counter- 
stained with methyl green). (F) Indirect IS-RT-PCR, five cycles, negative control — 
omission of reverse transcriptase. Note absence of staining (nuclei counterstained with 
methyl green). Even if nonspecific DNA repair occurs, the VDR probe will not bind to 
the repaired DNA and hence there is no signal in the nuclei. 


2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 215 

3.23. Reverse Transcription Step for Indirect IS-RT-PCR 

The technique of IS-RT-PCR, can be performed using the direct or indirect 
methods. Since the direct method is extremely prone to false-positive results 
when working on sections (Fig. 3), we recommend the use of the indirect 

1. Pre-hybridize the sections according to Subheading 3.13., steps 1-13 (steps 14- 
16 follow in Subheading 3.24.). 

2. Add 20 \iL random primer mix to each section and incubate at 80°C for 5 min. 

3. To each section, add 10 U of AMV RT; lx RT buffer; 10 mM DTT; 2 mM so- 
dium pyrophosphate; 15 U of RNasin; and 1 mM each of dATP, dCTP, dGTP, 
and dTTP in a final volume of 20 u,L, and incubate the samples at 42°C for 2 h. 

4. Wash twice in PBS or water. 

5. Fix in 0.4% paraformaldehyde-PBS for 20 min at 4°C. 

6. Dehydrate by incubating in 90% IMS for 5 min and allow to air-dry. 

3.24. PCR Step for Indirect IS-RT-PCR 

Annealing temperatures, Mg 2+ concentration, cycling parameters (typically 
5-10 cycles of amplication will be required; see Note 14) and extension times 
will vary with the particular mRNA under investigation. 

1. Place an Easiseal around each sample. 

2. Apply 50 [iL of amplification solution to each section and cover the Easiseal. 

3. Place the slides on the thermal cycler and start the reaction. 

4. When the PCR has finished, wash the slides twice in PBS or water. 

5. Transfer the slides to 0.4% paraformaldehyde-PBS and incubate for 20 min at 4°C. 

6. The remaining pretreatments described in Subheading 3.13., steps 14-16 are 
now performed. 

A comparison of radioactive and nonradioactive IS-RT-PCR is shown in 
Fig. 4. Indirect IS-RT-PCR in bone is shown in Fig. 5. 

3.25. Stringency Washes and Autoradiography for Indirect 

1. Denature the samples at 95°C for 5 min, using the Omnislide™, prior to the 
addition of probe. 

2. The hybridization conditions for IS-RT-PCR are exactly the same as those 
described for cDNA probes in Subheading 3.14. 

3. The post-hybridization stringency washes are as described for cDNA probes in 
Subheading 3.18. 

4. Autoradiography and counterstaining are performed as described in Subheading 3.21. 

15/Mee/199-224/F1 215 1 2/26/03, 10:47 AM 

I ■ I I I 


Mee and Hoyland 

1W* lt ? ? 

•v „* 

Fig. 4. Comparison of nonradioactive and radioactive indirect IS-RT-PCR. All sec- 
tions are after five cycles of IS-RT-PCR, and are kidney sections demonstrating VDR 
mRNA. (A) Nonradioactive detection with DIG. (B) Lightfield view showing silver 
grains over renal tubules. (C) Darkfield view to highlight the signal — note the absence 
of signal in the glomeruli (top right of B and C). 

Fig. 5. VDR mRNA expression in osteoclasts. (A) Normal ISH of Pagetic osteo- 
clast showing very low level of hybridization in this diseased tissue. (B) IS-RT-PCR, 
10 cycles, normal bone. (C) IS-RT-PCR, 10 cycles, Pagetic bone. (D) IS-RT-PCR, 
10 cycles, osteoclastoma. 

3.26. Direct IS-RT-PCR 

This procedure is not recommended for use on tissue sections — a variety of 
problems can cause false positive-results, but the most common problem is 
nonspecific DNA repair (Fig. 3). The protocol given below makes use of Tth 
enzyme (22), which possesses both RT and DNA polymerase activity and can 
therefore be used in a one-step protocol: 

1. Place Easiseal around sections. 

2. Pipet 50 u,L of reaction mixture onto each section. 

3. Perform amplification steps: 60°C for 30 min (RT step), followed by 20 cycles of 
amplification (as appropriate to mRNA under investigation), followed by a 
5 -min incubation at 60°C. 



2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 217 

3.27. Controls for ISH 

Both negative and positive controls are absolutely essential when perform- 
ing ISH. A specific reaction is of paramount importance. A false signal may 
arise from several causes, including sequence-independent binding of probe to 
nucleic acid; nonspecific binding to other tissue components; and artifacts of 
label visualization. 

A number of strategies can be used to assess the specificity of hybridization: 

1. Probe specificity: Check with Northern/Southern blots. 

2. Pretreatment of tissue with RNase or DNase. 

3. Hybridization with different fragments of specific sequence. 

4. Use of heterologous (unrelated) probe with similar GC content. 

5. Prehybridization of probe with specific cDNA or cRNA. 

6. Hybridization with nonspecific vector sequences. 

7. Reproducibility: Include tissue known to be positive or tissue with expressing 
and nonexpressing regions. 

8. Correlation with immunohistochemistry. 

9. If using riboprobes, a sense probe can be used as a further control. However, we 
and others have found that certain sense probes on certain tissues will produce 
positive staining. This staining is obviously not specific for the sequence being 
investigated by the antisense probe, hence many workers are now not using sense 
riboprobes as negative controls. 

3.28. Controls for IS-RT-PCR (Fig. 6) 

For IS-RT-PCR, the controls have to be even more rigorous. As with con- 
ventional ISH and IS-PCR a wide variety of control experiments must be per- 
formed to ensure that no false-negative or false-positive results are obtained. 

1. Use known negative and positive control samples. These must be used if avail- 
able to confirm that the reaction has occurred and that nonspecific amplification 
has not occurred. Another control within this category is to mix known numbers 
of positive and negative cells and check that the known negative cells show no 
positive signal following amplification. If more cells are apparently positive than 
should be, diffusion of amplification products has occurred. 

2. Use of RNase or DNase: These tests are performed when examining tissues for 
the respective nucleic acids. For example, if examining for DNA (such as exog- 
enous viral DNA) perform a DNase step to eliminate signal; if performing IS- 
RT-PCR, pretreat the sample with RNase to eliminate the signal. 

3. Omission of DNA polymerase. This tests for nonspecific DNA repair. Nicks 
and gaps can occur in genomic DNA (and in the DNA generated during the 
amplification step), especially in fixed, paraffin-embedded sections. Taq 
DNA polymerase will repair this damage during the extension reaction. When 
using direct IS-RT-PCR, this nonspecific action will lead to incorporation of 

15/Mee/199-224/F1 217 I 2/26/03, 10:47 AM 

I ■ I I I 



Mee and Hoyland 

m h 

Fig. 6. Controls for IS-RT-PCR. All samples have been hybridized with a VDR 
probe. (A) Negative control tissue — small cell lung cancer cell line (NCIH82) that 
shows no expression of VDR with any technique (including conventional RT-PCR). 
(B) Lightfield and (C) darkfield views of positive control tissue — kidney sections 
showing positive hybridization for VDR. (D) Negative control — RNase pretreatment 
prior to hybridization. (E) Negative control — no RT enzyme in RT step. (F) Negative 
control — no DNA polymerase in PCR. (G) Negative control — no primers in PCR. 

labeled nucleotides within the repaired DNA, which will produce a false-posi- 
tive result. A similar phenomenon can occur in cells that are undergoing 
necrosis or apoptosis. 

4. Omission of primers. This test will ensure the specificity of the primers, particu- 
larly if higher Mg 2+ concentrations have been used. The is also a further test to 
ensure that primer-independent nonspecific repair (as in the preceding) has not 
occurred. This control is particularly critical when performing direct IS-RT-PCR. 

5. Omission of RT enzyme when performing IS-RT-PCR. As with the use of RNase, 
this test will confirm that RNA, and not genomic DNA, is being amplified. 

6. Use of controls for the detection system. If using a nonradioactive detection sys- 
tem, these controls include omission of primary antibody. If using indirect IS- 
RT-PCR, a nonspecific probe (i.e., a probe that will not bind to the amplification 
product) can be used; alternatively, the probe can be left out altogether. A further 
test to ensure the specificity of the probe used for ISH is to perform the amplifi- 
cation reaction with a set of primers for a nucleic acid sequence that the probe 
should not hybridize with. 



2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 219 

7. Extraction of amplification products from the sample. The DNA can be extracted 
from the sample and the size of the amplification product verified by agarose gel 
electrophoresis. Be aware that, if direct amplification has been carried out, the 
products will appear slightly larger than expected owing to the presence of label 
within the DNA. 

4. Notes 

1 . Size of samples: If large samples are obtained then they should be cut into smaller 
pieces (less than 1-2 cm 3 ) prior to fixation. 

2. Duration of fixation: Samples of less than 1 mm thickness can be fixed in 2-3 h; 
tissues up to 1 cm thickness should be fixed for 12-24 h; larger samples should 
be fixed for up to 48 h. Precipitating fixatives such as ethanol and acetic acid 
provide better probe penetration, but they may be associated with loss of RNA 
(up to 75% losses have been quoted) and the tissue morphology is not so good. In 
view of this we do not use these fixatives for the detection of RNA. Crosslinking 
fixatives such as paraformaldehyde, formaldehyde, and glutaraldehyde generally 
provide better RNA retention and tissue morphology, but probe penetration is 
less good. The merits and drawbacks of the fixatives most commonly used are 
shown in the table below: 

Fixative Characteristics 

2% Glutaraldehyde Best RNA retention, probe penetration poor 

3% Formaldehyde Provides a good balance between RNA retention and 

probe penetration. No permeabilization needed for 

oligonucleotide probes 
4% Paraformaldehyde Provides a good starting point — usually no 

permeabilization steps needed 
10% Buffered formalin Used routinely in our laboratories 
Ethanol-acetic acid Unsuitable for RNA work 

3. Duration of decalcification: This is dependent on the sample size. Bone biopsies 
(8 mm trephine) usually take 10-14 d. Decalcification can be checked by placing 
the bone samples on top of a radiographic plate and exposing to X-rays for 5-8 sec at 
30-45 KvP. By developing the radiographic plate you will be able to assess 
whether any mineralization remains. Agitation of the specimens will increase the 
rate of decalcification. 

4. Embedding and processing tissues: Dehydration and embedding are best per- 
formed in a tissue processor. If this is not available in your own laboratory, 
embedding can be performed by most histopathology laboratories. 

5. Any calcified tissues that contain a large amount of cartilage (e.g., healing frac- 
ture callus or metaphyseal areas in growing bones) are extremely difficult to deal 
with, as the cartilaginous areas easily become detached from the slide. Once the 
sections have been mounted onto APES slides, incubating at 60°C for longer 
periods (up to 5 d) can help with this problem. 

15/Mee/199-224/F1 219 I 2/26/03, 10:47 AM 

I ■ I I I 


220 Mee and Hoyland 

6. Types of probe for ISH and IS-RT-PCR: Four types of probe are in common 
usage. These are double stranded complementary DNA (cDNA) probes; single 
stranded cDNA probes; RNA probes (riboprobe); and oligonucleotide probes. 
Each type of probe can be labeled using either radioactive or nonradioactive tech- 
niques. Radioactive labels include 32 P, 33 P, 35 S, and 3 H. Radioactive ISH gives 
more reproducible results and higher sensitivity than nonradioactive detection. 
Also, radioactive ISH is more readily quantifiable (silver grain counting). How- 
ever, the disadvantages include safety, waste disposal and reduced stability of 
labeled probe, prolonged time required for autoradiography and poor spatial reso- 
lution. Nonradioactive labels include biotin, DIG, and fluorescein. Nonradioac- 
tive probes are safer to use and have much greater stability; however, the 
protocols are not easily reproduced and the sensitivity of nonradioactive detec- 
tion is generally less than with radioactive probes, although many antibody 
detection systems have been developed that enhance the flexibility and sensitiv- 
ity of the method. 

7. Expected amounts of probe generated by DIG High Prime reaction: The product 
insert indicates that the DIG High Prime reaction yields approx 40 ng/mL of 
DIG-labeled DNA starting from 1 jxg of template after 1 h of incubation, we have 
found that it is more typical to have 20 ng/mL of DIG-labeled DNA following an 
overnight incubation. 

8. Size of riboprobes: RNA probes are most efficient when used as smaller frag- 
ments. The optimum size for ISH applications is generally 200-250 bases but 
should be determined empirically for each application. Probes <500 bp need not 
be hydrolyzed, but alkaline hydrolysis can be carried out for probes larger than 
this. The size of the probe can be checked by running an aliquot on a denaturing 
formaldehyde agarose gel. 

9. Prehybridization conditions: Note that for both ISH and IS-RT-PCR, for all new 
probes and tissues the optimum concentration and time of the preincubation 
should be determined empirically by trial and error. 

10. Negative control slides: Slides can be exposed to RNase during the 
prehybridization stage to provide a negative control. This tests specificity in pro- 
viding a check that the probe is binding to RNA, rather than other cellular com- 
ponents. During this incubation, the test slides (not RNase treated) should be 
incubated in 2x SSC. 

11. The immersion step in acetic anhydride-triethanolamine (see Subheading 3.13., 
step 14) can be extended up to 30 min to help reduce excessive background lev- 
els of signal (see Fig. 7). 

12. If background levels of staining are high, a prehybridization step can be included. 
This is merely an incubation for 60 min with all the components of the hybridiza- 
tion mix, except for the probe, performed at the normal temperature of hybridiza- 
tion (see Fig. 7). 

13. Temperatures for post-hybridization washes: This is based on the probe T m which 
depends on the length and G/C content of the probe. The T m is calculated using 
the following equation: 

15/Mee/199-224/F1 220 1 2/26/03, 10:47 AM 

I ■ I I I 

In Situ Hybridization and In Situ RT-PCR in Bone 



Fig. 7. Example of poor background staining and how to improve this. (A) 
Lightfield and (B) Darkfield views showing Bcl-2 expression in Pagetic bone. Note 
background signal over bone matrix and in clear areas within the marrow space. (C) 
Lightfield and (D) darkfield views showing the improvements seen by incorporating a 
prehybridization step and increasing the triethanolamine incubation to 30 min. Note a 
much cleaner signal over the osteoclasts. 

T m = 81.5°C + 16.61(log M) + 0.41(% GC) - (820/L) - 0.6% F - \-A% (mismatch) 

where F = formamide concentration, log M = log molar salt concentration, %GC 
= % of guanine and cytosine, and L = length of probe. 

When using cDNA probes, if high levels of background are still found, the 
formamide wash step can be performed at higher temperatures (up to 42°C) — 
this must be determined empirically for each probe and tissue. Similarly, the 
temperature of the 0.5 x SSC wash can also be increased to help reduce back- 
14. IS-RT-PCR is inefficient — we have shown that amplification is at best linear 
(compared with exponential for conventional tube RT-PCR) (16). This is impor- 
tant for two reasons; first, don't expect an exponential increase in signal, and 
second, criticism has been leveled at the technique owing to the fact that amplifi- 
cation has been used to generate a signal. The validity of results has hence been 
questioned by some authors owing to the "exponential increase" required to pro- 
duce a positive result (i.e. authors have questioned whether positive IS-RT-PCR 
findings are demonstrative of relevant levels of mRNA in vivo). However, the 


2/26/03, 10:47 AM 

I ■ I I I 


222 Mee and Hoyland 

increase is clearly not exponential, hence this criticism is unfounded. 
Too many cycles of IS-RT-PCR can actually reduce the level of signal. The rea- 
sons for this are unclear, although we have commented on the possibilities previ- 
ously (14). Nevertheless, when designing an IS-RT-PCR experiment, it is 
important to realize that only a few cycles of amplification will be needed (we 
have never gone above 10 cycles to get a positive result). This phenomenon 
appears to be related to the starting levels of mRNA, that is, the higher the origi- 
nal levels, the fewer cycles that will be needed; hence this phenomenon is espe- 
cially important in positive control tissues that are used to optimize protocols. 
(We have never used more than five cycles for positive control tissues.) 


1. Komminoth, P. and Long, A. A. (1993) In-situ polymerase chain reaction. An 
overview of methods, applications and limitations of a new molecular technique. 
Vir chows Arch. B 64, 67-73. 

2. Haase, A. T., Retzel, E. F., and Staskus, K. A. (1990) Amplification and detection 
of Lentiviral DNA inside cells. Proc. Natl. Acad. Sci. USA 87, 4971-4975. 

3. Nuovo, G. J., Gallery, F., MacConnell, P., Becker, J., and Bloch, W. (1991) An 
improved technique for in situ detection of DNA after polymerase chain reaction 
ampification. Am. J. of Pathol. 139, 1239-1244. 

4. Bagasra, O., Hauptman, S. P., Lishner, H. W., Sachs, M., and Pomerantz, R. J. 
(1992) Detection of human immunodeficiency virus type 1 provirus in mononuclear 
cells by in situ polymerase chain reaction. N. Engl. J. Med. 326, 1385-1391. 

5. Zehbe, I., Hacker, G. W., Rylabder, E., Sallstrom, J., and Wilander, E. (1992) 
Detection of single HPV copies in SiHa cells by in situ polymerase chain reaction 
combined with immunoperoxidase and immunogold-silver staining techniques. 
Anticancer Res. 12, 2165-2168. 

6. Komminoth, P., Long, A., Ray, R., and Wolfe, H. (1992) In situ polymerase chain 
reaction detection of viral DNA, single copy genes and gene rearrangements in 
cell suspensions and cytospins. Diagn. Mol. Pathol. 1, 85-97. 

7. Embleton, M. J., Gorochov, G., Jones, P. T., and Winter, G. (1992) In-cell PCR 
from mRNA: amplifying and linking the rearranged immunoglobulin heavy and 
light chain V-genes within single cells. Nucl. Acids Res. 20, 3831-3837. 

8. Heniford, B. W., Shum-Siu, A., Leonberger, M., and Hendler, F. J. (1993) Varia- 
tion in cellular EGF receptor mRNA expression demonstrated by in situ reverse 
transcriptase polymerase chain reaction. Nucl. Acids Res. 21, 3159-3166. 

9. Chen, R. H. and Fuggle, S. V. (1993) In situ cDNA polymerase chain reaction. A 
novel technique for detecting mRNA expression. Am. J. Pathol. 143, 1527-1534. 

10. Nuovo, G. J. (1994) Reverse transcriptase in situ PCR, in PCR In Situ Hybridiza- 
tion. Protocols and Applications, 2nd edit. Raven Press, New York, pp. 247-306. 

11. Patel, V. G., Shum-Siu, A., Heniford, B. W., Wieman, T. J., and Hendler, F. J. 
(1994) Detection of epidermal growth factor receptor mRNA in tissue sections from 
biopsy specimens using in situ polymerase chain reaction. Am. J. Pathol. 144, 7-14. 

12. Martinez, A., Miller, M. J., Quinn, K., Unsworth, E. J., Ebina, M., and Cuttitta, F. 

15/Mee/199-224/F1 222 I 2/26/03, 10:47 AM 

I ■ I I I 


In Situ Hybridization and In Situ RT-PCR in Bone 223 

(1995) Non-radioactive localization of nucleic acids by direct in situ PCR and in 
situ RT-PCR in paraffin-embedded sections. /. Histochem. Cytochem. 43, 119-1 '47 '. 

13. Martinez, A., Miller, M. J., Unsworth, E. J., Stegfried, J. M., and Cuttitta, F. (1995) 
Expression of adrenomedullin in normal human lung and in pulmonary tumours. 
Endocrinology 136, 4099-4105. 

14. Mee, A. P., Davenport, L. K., Hoyland, J. A., Davies, M., and Mawer, E. B. (1996) 
Novel and sensitive detection systems for the vitamin D receptor — in situ-reverse 
transcriptase-polymerase chain reaction and immunogold cytochemistry. /. Mol. 
Endocrinol. 16, 183-195. 

15. Mee, A. P., Hoyland, J. A., Braidman, I. P., Freemont, A. J., Davies, M., and 
Mawer, E. B. (1996) Demonstration of vitamin D receptor transcripts in actively 
resorbing osteoclasts in bone sections. Bone 18, 295-299. 

16. Mee, A. P., Denton, J., Hoyland, J. A., Davies, M., and Mawer, E. B. (1997) 
Quantification of vitamin D receptor mRNA in tissue sections demonstrates the 
relative limitations of in situ-reverse transcriptase-polymerase chain reaction. /. 
Pathol. 182, 22-28. 

17. Hoyland, J. A., Mee, A. P., Band, P., Braidman, I. P., Mawer, E. B., and Freemont, 

A. J. (1997) Demonstration of oestrogen receptor mRNA in bone using in situ- 
reverse transcriptase-polymerase chain reaction. Bone 20, 87-92. 

18. Mee, A. P., Dixon, J. A., Hoyland, J. A., Davies, M., Selby, P. L., and Mawer, E. 

B. (1998) Detection of canine distemper virus in 100% of Paget's disease samples 
by in situ-reverse transcriptase-polymerase chain reaction. Bone 23, 171-175. 

19. Long, A. A., Komminofh, P., Lee, E., and Wolfe, H. J. (1993) Comparison of 
indirect and direct in situ polymerase chain reaction in cell preparations and tissue 
sections. Histochemistry 99, 151-162. 

20. Myers, T. W. and Gelfand, D. H. (1991) Reverse transcription and DNA amplifi- 
cation by a Thermus thermophilus DNA polymerase. Biochemistry 30, 7661. 

21. Walsh, L., Freemont, A. J., and Hoyland, J. A. (1993) The effect of tissue decalci- 
fication on mRNA retention within bone for in situ hybridization studies. Int. J. 
Exp. Pathol. 7 '4, 237-241. 

22. Rentrop, M., Knapp, B., Winter, H., and Schweizer, J. (1986) Aminoalkylsilane- 
treated slides as support for in situ hybridization of keratin cDNAs to frozen sec- 
tions under varying fixation and pretreatment conditions. Histochem. J. 18, 271-276. 

15/Mee/199-224/F1 223 1 2/26/03, 10:47 AM 

I ■ I I I 



2/26/03, 10:47 AM 

I ■ I I I 



Techniques for the Study of Apoptosis in Bone 

Brendon S. Noble and Hazel Y. Stevens 

1 . Introduction 

Over the past five years there has been an explosion of interest in the pro- 
cess of apoptotic cell death. The fact that, in contrast to necrotic death, 
apoptosis constitutes a precisely regulated, energy-dependent form of death 
has led to interest in the potential to control this death process through an 
understanding of the molecular mechanisms involved in its initiation and 
occurrence. Early descriptions of apoptosis were based on morphological 
changes in cells (cell shrinkage, condensation and margination of chromatin 
nuclear fragmentation, and production of membrane-bound apoptotic bodies), 
and despite the more recent identification of a number of biochemical (e.g., 
ordered DNA fragmentation and activation of a range of caspase enzymes) and 
ultrastructural phenomena (e.g., externalization of phosphatidylserine mol- 
ecules at the plasma membrane) specific to apoptosis, the morphological crite- 
ria are still often regarded as the "gold standard" (1). Unfortunately, the use of 
morphological criteria to assess the apoptotic state in many cell types involves 
the use of high magnification microscopy (often electron microscopy), which 
precludes studying the high numbers of individual cells necessary for an as- 
sessment of apoptosis at a tissue level. We have attempted to address these 
difficulties by identifying more than one apoptotic characteristic at relatively 
low magnifications. Hence we assess (1) DNA fragmentation in situ, (2) cell 
loss, and (3) identification of internucleosomal sized increments of DNA, pro- 
duced during DNA destruction (DNA ladders). The loss of cells through the 
energy-dependent, controlled mechanism of apoptosis has wide-ranging influ- 
ence on the function of all body tissues and bone is no exception. The loss of 
bone cells by apoptosis occurs as an entirely normal component of the life- 
time repertoire of these cells. It is thought to control the number of functional 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


16/Noble/225-236/F1 225 1 2/26/03, 10:47 AM 

I ■ I I I 


226 Noble and Stevens 

osteoblasts at the bone surface (2), and to play a role in endochondral bone 
formation by chondrocytes (3); in addition, it deletes osteoclasts at the end of 
their resorption work cycle (4). It has been suggested that in the osteocyte 
apoptotic death might play a role in the load and microdamage engendered 
targeting of bone resorption (5). The importance of apoptosis in normal skel- 
etal physiology is emphasized by recent evidence that apoptosis is perturbed or 
deregulated in a number of disease situations in bone (6,7). 

The study of cellular morphology in mineralized tissues is complicated by 
difficulties experienced in cutting thin sections of the tough, brittle mineral 
component of the bone and by the contrasting material properties of the miner- 
alized bone and nonmineralized components, such as bone marrow and carti- 
lage which are present within the same tissue. Sectioning of bone can be 
facilitated by the use of plasticized embedding materials such as 
methylmethacrylate to support the heterogeneous materials within bone tissue 
(see the chapter by van Leeuwen and Derkx, this volume). Unfortunately, the 
use of plastic embedding materials greatly hinders penetration of the tissue 
sections by enzymes, antibodies, and various other reagents necessary to detect 
apoptosis. Removal of embedding material with a deplasticizing step often 
results in loss of the small fragments of DNA associated with apoptosis, result- 
ing in underestimates of apoptosis. Ideally, fresh frozen cryosections should be 
used to maximize the chances of identifying and quantifying enzymes or anti- 
gens in bone cells, as these methods avoid masking or changing conformation 
of the molecules studied. Using a tungsten carbide edged knife and a heavy- 
duty cryostat (ideally motorized Bright cryostat; see the chapter by Bord, this 
volume), it is possible to cut cryosections of bone. 

2. Materials 

2. 1. Toluidine Blue Staining 

1 . Picric formalin: Mix 60 mL of formalin (40% aqueous formaldehyde), 500 mL of 
95% ethanol, 40 mL of glacial acetic acid, and 400 mL of distilled water. The 
solution keeps for about 4 mo at room temperature (RT). Make a fresh batch as 
soon as a cloudy precipitate forms or the solution smells clearly of formic acid. 

2. Toluidine blue (Gurr, Poole, UK). 

3. /j-Butyl alcohol (Sigma, Dorset, UK). 

4. Citifluor (glycerol-PBS mix; Agar Scientific, Essex, UK). 

5. Light green (Gurr). 

2.2. LDH Assay 

1. Gly-Gly powder (Sigma). 

2. Polypep (40%). Weigh out 1.32 g Gly-Gly powder, add 100 mL of distilled water, 
then adjust pH to 8.0 by adding approx 34 mL of 0. 1 M NaOH. Then add 80 g of 

16/Noble/225-236/F1 226 1 2/26/03, 10:47 AM 

I ■ I I I 


Apoptosis in Bone 227 

Polypep (Sigma). Leave stirring overnight. Make up to 200 mL with distilled 
water and store at 4°C (solidified stock solution keeps for 6-12 mo at this tem- 

3. Lactic acid (Sigma). 

4. Nicotinamide adenine dinucleotide (NAD; BDH/Merk, Poole, Dorset, UK). 

5. Nitroblue tetrazolium (NBT; Sigma). 

2.3. Nick Translation 

1. Paraformaldehyde (Sigma) dissolved in PBS (4% w/v) by gentle heating to 60°C. 
Caution: Do not boil! Ideally make fresh on day and do not shake before use. 

2. Phosphate-buffered saline (PBS; Sigma). 

3. Ethylenediaminetetraacetic acid (EDTA; BDH) 0.25 M in 50 mM Tris-HCl, pH 
7.4 (stable for 1 yr). 

4. Nucleotides: 100 mM of dGTP Li salt (cat. no. 1051 466); 100 mM of dATP Li 
salt (cat. no. 1051 440); 100 mM of dCTP Li salt (cat. no. 1051 458) (all pur- 
chased from Roche Diagnostics). With all Roche products do not use beyond 
expiration date. 

5. Polymerase buffer: 6.055g/L of 50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 (238 mg/ 
500 mL), 0. 1 mM dithiothreitol (7.7 mg/500 mL)]. Stable for 4 mo stored at 4°C. 

6. Digoxigenin (DIG)-ll-dUTP 25 uL of 25 nmol (Roche, cat. no. 157 3152). 

7. 250 U (50 u,L) of DNA polymerase 1 (Roche, cat. no. 642 711). 

8. Anti-DIG-fluorescein (FITC) Fab fragments (200 fig) (Roche, cat. no. 1 207 741). 

9. Normal sheep serum (Sigma). 

10. Positive control DNase 1 (Sigma, cat. no. D4263). Do not use after expiration date. 

11. Propidium iodide (PI; Sigma). Make up a stock solution of 100 \ig/mh in dis- 
tilled water. Stable for 6 mo at 4°C, stored in the dark. The working solution is 
10 pL of PI stock solution in 10 mL of distilled water. 

2.4. DNA Gel Electrophoresis 

1. Reagent B, sodium perchlorate, Nucleon silica (Nucleon Biosciences, Scotlab, 
Glasgow, UK). Do not use after expiration date. 

2. Ribonuclease A (Sigma). 

3. Chloroform (BDH), incubated at -20°C for the purification step. 

4. TBE: 54 g of Tris base, 27.5 g of boric acid, 20 mL of 0.5 M EDTA pH 8.0. 
Stable for 6 mo at 4°C. 

5. Agarose (gel electrophoresis grade; Gibco, Paisley, Scotland). 

6. 6X Gel loading buffer (stored refrigerated at 4°C, stable for 6 mo): 0.025 g of 
bromophenol blue, 0.025 g of xylene cyanol FF, 3 mL of glycerol, 7 mL of water. 
All reagents purchased from Sigma. A commercially available loading buffer 
can be substituted. 

7. 100-Basepair (bp) ladder (Gibco). Do not use after expiration date. 

8. Horizontal gel electrophoresis tank (Bio-Rad) and power supply delivering up to 
150 V. 

9. UV transiiluminator. 

16/Noble/225-236/F1 227 I 2/26/03, 10:47 AM 

I ■ I I I 


228 Noble and Stevens 

3. Methods 

3. 1. Morphological Evidence of Apoptosis 
by Toluidine Blue Staining 

Apoptotic cells can be identified in sections of bone by toluidine blue stain- 
ing. The procedure stains the nuclei blue and enables visualization of conden- 
sation, blebbing, or fragmentation of the nucleus prior to packaging of nuclear 
and cytoplasmic contents into apoptotic bodies. 

1. Fix cryostat sections in picric-formalin for 10 min at RT. 

2. Make up a 0.1% solution of toluidine blue (0.2 g in 200 mL of H 2 0). 

3. Stain the sections for 30 min. 

4. Blot dry and place in buffer for resin sections or /!-Butyl alcohol for frozen sec- 
tions (see Note 1) for 2 min. 

5. If a counterstain is required, use 1% light green in distilled H 2 for 2 min and 
rinse again with distilled water. 

6. Mount in Citifluor. (The standard technique involves dehydration and mounting 
in DePeX which may or may not suit the material being stained. In general, bone 
material does not require this dehydration step.) 

3.2. Cell Viability by LDH Assay 

During most of the apoptotic time course cells maintain intact cell membranes 
and active metabolic processes. This is in distinct contrast to the necrotic death 
process in which cell membranes rupture and cellular activities rapidly decline. 
Hence, detection of DNA fragmentation in cells with intact membranes and active 
metabolic enzymes will indicate apoptosis rather than necrosis. Loss of cell viabil- 
ity also represents the final "outcome" of the apoptotic process and, in the case of 
the osteocytes, entombed within lacunae in bone, it will indicate the site of cell 
loss. To determine cell viability in cryosections we use histochemical detection of 
LDH enzyme activity at the cellular level. Rather than attempt to measure enzyme 
activities quantitatively, the technique is designed to be highly sensitive to enzyme 
activity to ensure that any active enzyme present is identified. 

Because apoptosis occurs only in living cells it is important to use the LDH assay 
alongside DNA ladders to test which cells were alive when the tissue was prepared 
for sectioning. The method given is a modification of the methods of Wong et al. 
1987 (6) and Farquharson et al. 1992 (8). Purple staining indicates viable cells; 
absence of staining implies a dead cell or presence of an empty lacuna (see Fig. 1). 

1. Generate cryostat sections (10 u,m thick) on microscope slides and keep slides at 
-20°C until use. 

2. When ready, defrost slides at RT for a few minutes. 

3. Melt 40% Polypep in ajar surrounded by hot water and to 10 mL of Polypep add 44 u.L 
of lactic acid (final cone. 60 mM), 17.5 mg of NAD (final cone. 1.75 mg/mL), then 
adjust pH to 8.0 with cone. NaOH before adding 30 mg NBT (final cone. 30 mg/mL). 

16/Noble/225-236/F1 228 1 2/26/03, 10:47 AM 

I ■ I I I 

Apoptosis in Bone 



Glass slide 

Fig. 1. Cell viability determined in situ using LDH activity as a marker. Cells in 
frozen sections are stained for lactate dehydrogenase activity and examined using light 
microscopy. (A) Live osteocytes stained (dark) for active lactate dehydrogenase. 
Arrows illustrate two examples of live cells. (B) Region of bone containing dead 
osteocytes showing no staining for LDH activity. Cells on bone surface stain positive. 
(C) Diagram illustrates the use of plastic rings for LDH staining. The reaction mix is 
placed in a plastic ring sealed at the base and top with vaseline to allow prolonged 
incubation at 37 C C. 

4. Adhere plastic rings (made from slices of polymethylmethacrylate [PMMA] tub- 
ing) over the specimens onto the slides using Vaseline, add the reaction mix into 
the ring excluding air bubbles and place coverslips on top to minimize evaporation. 



2/26/03, 10:47 AM 

I ■ I I I 


230 Noble and Stevens 

5. Incubate for 3 h at 37°C in a humidified chamber. 

6. Rinse in warm water (approx 50°C) to remove reaction medium. 

7. Rinse in acetone (30 s) to remove soluble pink formazan. 

8. Rinse in PBS and mount in Citifluor. Coverslips can be sealed with clear nail 
varnish if needed. 

3.3. Nick Translation 

In general, cells undergoing apoptosis cleave their DNA into 
internucleosomal sized increments (increments of 180-200 bp) so that, when 
the DNA from apoptotic cells is electrophoresed on an agarose gel, the DNA 
presents itself at regular intervals in the run giving the appearance of a "DNA 
ladder" (see Subheading 3.4.)- The level of DNA fragmentation that occurs 
during apoptosis is extremely high, vastly in excess of that associated with 
normal DNA repair or during necrosis. The in situ nick translation technique 
described here uses the DNA polymerase I (Klenow fragment) driven incor- 
poration of DIG-conjugated nucleotides into DNA strand breaks to identify 
cells in culture or in tissue sections that contain large amounts of highly frag- 
mented DNA characteristic of apoptotic cells (see Fig. 2). The technique has 
been purposely designed to be relatively insensitive with regard to DNA frag- 
ment identification so that only cells with massive amounts of fragmented 
DNA will be identified. In this way the technique has an increased specific- 
ity for apoptotic cells (7). We have found that when used on bone sections 
the technique provides a more accurate and consistent method of identifica- 
tion of apoptosis that the more commonly used TUNEL staining, which 
employs terminal deoxynucleotide transferase (TdT). The reasons for this 
are unknown but likely include the fact that TUNEL methods greatly amplify 
the fragmentation signal due to addition of multiple labeled (biotinylated) 
nucleotides at the 3 'OH termini of a break (in a chain reaction) while nick 
translation applies only a single set of nucleotides. In addition, commercial 
kits for TUNEL often include a harsh proteinase K digestion step which prob- 
ably introduces false positives. 

While it is our experience that all bone cells tested so far (from various 
species) undergo this type of internucleosomal DNA fragmentation during 
apoptosis it is wise to always run a positive control (in which apoptosis has 
been positively identified by other means) to test the applicability of this tech- 
nique to the specific tissue of interest. 

DNA is susceptible to degradation by endonucleases. This means that all 
vessels, tips, and solutions should be sterile before use. In practice, the use of 
autoclaved microcentrifuge tubes, the wearing of gloves and sterile filtered 
stock solutions of buffer are usually sufficient to exclude exogenous endonu- 
clease activity and enable the unambiguous use of the technique. 

16/Noble/225-236/F1 230 1 2/26/03, 10:47 AM 

I ■ I I I 


Apoptosis in Bone 


Fig. 2. Cells containing large amounts of highly fragmented DNA are labeled using 
a nick translation technique. Fragmented DNA in osteocytes resident in bone is iden- 
tified after incorporation of labeled nucleotides using a nick translation reaction. (A) 
Propidium iodide staining of osteocyte nuclei. (B) Apoptotic osteocytes labeled posi- 
tive for fragmented DNA (FITC). Arrows denote two example cells positive for frag- 
mented DNA. (C) PI staining of osteocyte nuclei in the negative control (no 
polymerase enzyme). (D) Negative control (no polymerase enzyme) showing lack of 
FITC-labeled cells. 

3.3.1. Preparing the Sections/Cells 

1. Prepare fresh 7-p.m cryostat sections of rat tibia or prepare bone cell cultures {see 
Note 2). 

2. Fix in 4% paraformaldehyde in PBS for 10 min at RT. 

3. Wash three times in PBS. 

4. For sections, decalcify in 0.25 M EDTA in 50 mM Tris-HCl, pH 7.4, for 10 min. 
Cell cultures do not require any decalcification. 

5. Wash three times in PBS. 

6. Thoroughly air dry. Store at 4°C until ready to use. 



2/26/03, 10:47 AM 

I ■ I I I 


232 Noble and Stevens 

3.3.2. Preparing the Nick Translation Mixture 

The following recipe is based on an allocation of 50 pL of reaction mix per 
slide and is suitable for five slides plus a control slide (without polymerase 

1 . Make up stock solutions of dGTP, dATP, and dCTP in three separate Eppendorf 
tubes by adding 1 \ih of dGTP, dATP, and dCTP (from the 100 mmol/L stock) to 
300 [iL of polymerase buffer, resulting in working stocks of 0.3 mmol/L for each 

2. Then add 3 \iL of each of the working stocks of nucleotides into 300 \ih of poly- 
merase buffer (= polymerase mix, containing 3 \xM of each of nucleotides). 

3. Add 1 [iL of DIG-1 1-dUTP into the polymerase mix. 

4. Take out 50 p,L of polymerase mix for the negative control. 

5. Then add 1.25 [iL of DNA polymerase to the remaining 250 jxL of polymerase 
mix (= nick translation mix) and vortex-mix well. 

6. Use DNase I at 0.2 mg/mL in PBS as a positive control in a subset of sections 
(see Subheading 3.3.3.). 

3.3.3. Nick Translation Technique 

1. Treat one group of sections with DNase (positive control) for 1 h at 37°C to 
produce breaks in the DNA. 

2. Incubate all other sections/cells in nick translation mix for 1 h at 37°C in a humidified 
chamber. Remember negative controls have no DNA polymerase in the mix. 

3. Wash three times in PBS and keep moist at all times. 

4. Incubate in 1:8 anti-DIG FITC Fab (300 uT of PBS, 37.5 ^L of anti-DIG FITC 
Fab, 15 [iL of sheep serum) for 1 h at RT. 

5. Wash three times in PBS. 

6. Counterstain in PI (10 \iL of PI stock solution in 10 mL of distilled water) for 30 s 
(cells) and 2 min (sections). 

7. Wash thoroughly in distilled water. 

8. Mount in Citifluor, keep dark. 

9. The sections/cells are examined under a fluorescent microscope to distinguish 
FITC stained (green) apoptotic cells from all cells (stained red with PI). Positive 
controls show a large number of cells (often approaching 100% of cells) with 
fragmented DNA (FITC positive) and negative controls should not show any 
FITC-positive cells, just a low background fluorescence of the bone. 

3.4. DNA Ladders 

Breaking DNA into internucleosomal sized increments during apoptosis 
(increments of 180-200 bp) leads to the production of a "DNA ladder" when 
the DNA is electrophoresed on an agarose gel (9). A small number of cell types 
produce DNA fragments during apoptosis that are much larger (200-300 kbp 
and 30-50 kbp) and hence do not provide this ladder pattern (10). It is thought 

16/Noble/225-236/F1 232 I 2/26/03, 10:47 AM 

I ■ I I I 


Apoptosis in Bone 233 

that the larger fragments are produced as a prelude to the production of 
oligonucleosomal fragments and that apoptotic cells not showing DNA ladders 
have stopped DNA fragmentation at this earlier stage in the process. It is pos- 
sible to identify the larger fragments using pulse field electrophoresis (not cov- 
ered in this chapter). 

While it is our experience that all bone cells tested so far (from various 
species) undergo internucleosomal DNA fragmentation during apoptosis, it is 
again wise to always include a positive control (in which apoptosis has been 
positively induced, e.g., heating cell culture flask at 44°C for 30 min, or by 
using sections of material previously tested by nick translation) for the cell 
type under investigation to test the fragmentation pattern (see Fig. 3 and Note 
4). As for the nick translation assay, use autoclaved microcentrifuge tubes, 
wear gloves and use sterile filtered stock solutions of buffer to exclude exog- 
enous endo nuclease activity. 

3.4.1. Preparation of Cells and Tissue 

1. For a confluent T75 flask of bone cells (approx 4 x 10 6 cells), aspirate medium 
and wash the monolayer gently in PBS, aspirate to dryness, and place in freezer 
(-80°C) immediately. 

2. Thaw by adding 1 mL of PBS and scrape cells off into a 1.5- to 2-mL Eppendorf 

3. Centrifuge at 600g, 4°C, 5 min. 

4. For sections generated on a cryostat (15-20 at 10 jim thick; undecalcified tissue), 
collect these directly into a bijou/Eppendorf tube and store immediately at -80 C C. 
Thaw when ready to use. 

3.4.2. Isolation of DNA 

The isolation of DNA uses a commercial kit, supplied by Nucleon Bio- 
sciences (see Note 3) and our protocol has been published in their scientific 

1 . To the cell pellet or sections add 340 \xL of Reagent B (SDS based). Vortex-mix. 
Leave for 40 min. 

2. Centrifuge at 600g for 5 min. Decant the supernatant to another tube. 

3. Digest RNA (optional) by addition of 2.5 [iL of ribonuclease A (Sigma) to each 
tube (final cone. 50 u,g/mL) and incubate for 30 min at RT. 

4. Remove protein by adding 100 jxL of sodium perchlorate per tube and mixing on 
a rotary mixer at 37°C for 20 min, followed by 20 min at 65°C. 

5. Add 580 \ih of chloroform (stored at -20°C). Place on rotary mixer for 20 min at 

6. Transfer the contents to a 2-mL Nucleon /Eppendorf tube. Nucleon tubes have a 
plastic division which makes separation of the organic and aqueous layers easier. 
Centrifuge at 1300g for 1 min. 

16/Noble/225-236/F1 233 1 2/26/03, 10:47 AM 

I ■ I I I 


Noble and Stevens 



Fig. 3. DNA ladders indicative of apoptosis. DNA from apoptotic cells produces 
multiple bands of approx 180-bp increments when run on an agarose gel. Lanes A, B, 
DNA from apoptotic cells producing the characteristic "ladder" pattern. Individual 
bands are highlighted with arrows. Lane C, 1000-bp markers. 

7. Add 45 \ih of Nucleon silica suspension, then centrifuge at 1300g for 4 min. 

8. Pour off (upper, aqueous) DNA phase into a fresh tube. 

9. Centrifuge at 1300g for 30 s to pellet any silica. 

10. Transfer supernatant to a fresh tube. 

11. To precipitate DNA add 880 u.L of absolute ethanol and invert the tube to mix. 

12. Centrifuge at 4000g for 5 min to pellet DNA, discard the supernatant (or, optional, 
add 100 [iL 3 M sodium acetate to the ethanol waste and leave at -20°C overnight 
to precipitate further DNA). 



2/26/03, 10:47 AM 

I ■ I I I 


Apoptosis in Bone 235 

13. Wash the DNA pellet by adding 1 mL of 70% ethanol in distilled water at 4°C 
and place on a rotary mixer at RT for 20 min. 

14. Centrifuge briefly at 4000g for 5 min. Remove ethanol and leave the pellet to air- 
dry (not to complete dryness). The samples are usually solubilized in TBE and 
run the same day. DNA can be stored in 70% ethanol at -20°C so that the proce- 
dure can be interrupted at step 12. 

3.4.3 Agarose Gel Electrophoresis 

1. Dissolve 1.5% (w/v) high-purity agarose in TBE (100 mL of IX TBE for a 70-mL 
gel) in a 250-mL glass flask by microwaving at a medium wattage (500 W) for 
approx 3 min. To avoid overheating, take out the flask after 2 min with tongs and 
swirl the gel. 

2. Pour the gel into a suitable container and let set for 45 min. Then hydrate gel with TBE. 

3. Add 20 [iL of TBE to pellets, incubate at 65°C for 3 min, cool, and add 4 uL of 
loading buffer (LB 6x). 

4. Load approx 10 |xg of DNA per lane, running approx 20 uL in each lane (see 
Note 3) 

5. Include one lane of a 100 bp-standard. 

6. Run the gel at 20 V overnight, until the front marker is near the end of gel (see 
Note 4). 

7. Soak the gel in 300 mL of 1 x TBE with 15 [iL of ethidium bromide (final concen- 
tration 0.5 jig/mL) for 45 min. 

8. Visualize using UV Transilluminator (see Note 5). 

4. Notes 

1. Resin-embedded sections should be washed in buffer rather than in 77-butyl alco- 
hol after toluidine blue staining to avoid shrinkage and crinkling of the section. 

2. It is crucially important for the nick translation technique always to use freshly 
cut and fixed sections. Storage of frozen sections induces damage of DNA upon 
defrost and results in false-positive results. 

3. We have most experience with the Nucleon kit; however, any commercial kit to 
isolate small quantities of DNA would probably suffice. 

4. The amount of DNA that may be loaded on a gel depends on several factors: 

a. Well volume. 

b. Fragment size. The capacity of the gel drops sharply as the fragment size 
increases, especially over a few kilobases. 

c. Distribution of fragment sizes 

d. Voltage gradient. Higher voltage gradients are better suited to DNA frag- 
ments under 1 kb; lower voltages are better suited to fragments over 1 kb. 

The least amount of DNA in a single band that can be reliably detected with 
ethidium bromide is approx 10 ng and about 60 pg with SYBR® Green I stain. 
The maximum amount of DNA that can be run as a sharp, clean band is about 100 ng. 
Overloaded DNA results in trailing and smearing, a problem that will become 
more severe as the size of DNA increases. 

16/Noble/225-236/F1 235 1 2/26/03, 10:47 AM 

I ■ I I I 


236 Noble and Stevens 

5. Using standard agarose gel electrophoresis, the presence of a ladder is a clear 
indication of apoptosis but the absence of a ladder is not proof that apoptosis has 
not occurred until you have established the way in which DNA fragmentation 
occurs in the cell type under investigation. In addition, it is important to verify 
the apoptotic state using multiple criteria if possible. 


1. Wyllie, A. H., Kerr, J. F. R., and Currie, A. R. (1980) Cell death: the significance 
of apoptosis. Int. Rev. Cytol. 68, 251-306. 

2. Jilka, R. L., Weinstein, R. S., Bellido, T., Roberson, P., Parfitt, A. M., and 
Manolagas, S. C. (1999) Increased bone formation by prevention of osteoblast 
apoptosis with parathyroid hormone. /. Clin. Invest. 104, 439-446. 

3. Stevens, H.Y., Reeve, J., and Noble, B.S. (2000) Bcl-2, tissue transglutaminase 
and p53 protein expression in the apoptotic cascade in ribs of premature infants. /. 
Anat. 196, 181-191. 

4. Kameda, T., Ishikawa, H., and Tsutsui, T. (1995) Detection and characterisation of 
apoptosis in osteoclasts in vitro. Biochem. Biophys. Res. Commun. 207, 753-760. 

5. Verborgt, O., Gibson, G. J., and Schaffler, M. B. (2000) Loss of osteocyte integ- 
rity in association with microdamage and bone remodelling after fatigue in vivo. 
/. Bone Miner. Res. 15, 60-67. 

6. Wong, S. Y., Evans, R. A., Needs, C, Dunstan, R., Hills, E., and Garvan, J. (1987) 
The pathogenesis of osteoarthritis of the hip. Evidence for primary osteocyte 
death. Clin. Orthop. 214 305-312. 

7. Noble, B.S., Stevens, H., Loveridge, N., and Reeve, J. (1997) Identification of 
apoptotic changes in osteocytes in normal and pathological human bone. Bone 20, 

8. Farquharson, C., Whitehead, C, Rennie, S., Thorp, B., and Loveridge, N. (1992) 
Cell proliferation and enzyme activities associated with the development of avian 
tibial dyschondroplasia: an in situ biochemical study. Bone 13, 59-67. 

9. Wyllie, A. H. (1980) Glucocorticoid-induced thymocyte apoptosis is associated 
with endogenous endonuclease activation. Nature 284, 555-556. 

10. Oberhammer, F., Wilson, J. W., Dive, C, et al. (1993) Apoptotic death in epithe- 
lial cells: Cleavage of DNA to 300 and/or 50 kb fragments prior to or in the 
absence of internucleosomal fragmentation. EMBO J. 12, 3679-3684. 

16/Noble/225-236/F1 236 1 2/26/03, 10:47 AM 

I ■ I I I 



Protein Localization in Wax-Embedded 
and Frozen Sections of Bone Using 

Sharyn Bord 

1 . Introduction 

Immunohistochemistry can provide valuable information regarding protein 
expression in different cell types at specific stages of differentiation during 
bone modeling and remodeling. By combining immunohistochemistry with 
other techniques, it is possible for the researcher to determine protein expres- 
sion, in relation to mRNA production, enzyme activity and bone remodeling, 
on the same sample of bone. This chapter covers the localization of protein in 
human bone by immunohistochemistry using an indirect immunoperoxidase 
method, and considers both frozen and wax-embedded sections. Methods for 
immunostaining of plastic embedded tissue can be found in the chapter by Van 
Leeuwen and Derkx, this volume. Immunohistochemistry is based on incubat- 
ing high-affinity antibodies on tissue sections to detect patterns of expression 
for specific antigens within the tissue. This can be achieved using either an 
immunofluorescence-based technique in which the antibody is conjugated with 
a fluorochrome, or an enzyme-labeled antibody method. The latter procedure 
has several advantages; it utilizes the same enzyme complexes for all primary 
antisera irrespective of its origin in different animal species; stained sections 
may be permanently mounted; and sections can be viewed using brightfield 
microscopy. Detection methods have increased the sensitivity of immunohis- 
tochemistry. There are now several ways that the signal of antibody-antigen 
binding can be amplified, thus allowing the use of only small amounts of anti- 
body. The method of choice for this chapter is the avidin-biotinylated enzyme 
complex (ABC) method first described by Hsu et al. (1) for localization of 
antigens in the thyroid. Primary antibody binds to specific epitopes of the anti- 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


17/Bord/237-248/F1 237 I 2/26/03, 10:47 AM 

I ■ I I I 





primary antibody binding to antigen 


biotinylated second antibody binds to primary antibody 

ABC amplifies sites of antigenicity 

signal detected with enzyme substrate 


primary and secondary antibody ^f Avidin Biotin Complex 


Fig. 1. Steps involved in the indirect immunoperoxidase system using an ABC 
amplification stage. (A) After fixation and blocking, the primary antibody is applied 
to the section. This binds specifically to the antigen. (B) A biotin-labeled second anti- 
body binds to the primary antibody and introduces many biotins at this binding site. 
(C) The ABC enzyme binds to the biotinylated sites, thus increasing the signal. (D) 
The substrate reacts with the enzyme complex and gives a color reaction at sites of 

gen within the section. A biotin-labeled secondary antibody is added that intro- 
duces many biotins at the sites of primary antibody binding. The next layer, 
ABC, binds to the biotin-labeled second antibody, amplifying the initial signal 
which is then visualized by addition of a substrate and chromogen to the en- 
zyme. This produces a color reaction at sites of positive expression that can be 
assessed using brightfield microscopy (Fig. 1). We have successfully used this 
technique to detect various proteins in bone (2,3). 

2. Materials 

Most general reagents can be obtained from Sigma. 

1. Phosphate-buffered saline (PBS) 5x stock: Add 42.6 g of NaCl, 16.08 g of 
Na 2 HP04-2H 2 0, and 0.78 g of NaH 2 P0 4 to 900 mL distilled water. Adjust the 
pH to 7.4 with 6 N HC1 and make up to 1 L. Store at room temperature and dilute 
1: 5 with distilled water before use. 

2. Buffered EDTA: Add 72.5 g of ethylenediamine tetraacetic acid disodium salt 
(EDTA) to 500 mL of distilled water and adjust to pH 7.4-7.6 with NaOH. Add 
five PBS tablets to the solution and mix thoroughly using stirrer. 



2/26/03, 10:47 AM 

I ■ I I I 


Protein Immunolocalization in Bone 239 

3. Bovine serum albumin (BSA): Dissolve 1 mg of high-grade crystalline BSA in 1 mL 
of PBS to give a 0.1% solution. 

4. Neutral buffered formalin (NBF): Add 5.2 g of NaH 2 P0 4 -2H 2 and 6.4 g of 
Na 2 HP0 4 (anhydrous) to 97 mL of 37% formaldehyde. Combine with in 900 mL 
of distilled water and adjust the pH to 7.2 with 6 N HC1. Make up to 1 L with 
distilled water. 

5. 4% Paraformaldehyde fixative: Add 4 g of paraformaldehyde to 50 mL of PBS. 
Add a few drops of 2.5 M NaOH and stir on a hot plate in a fume hood until clear. 
Allow to cool and add 40 mL of PBS. Adjust pH to 7.3 with 1 N acetic acid. Make up 
to 100 mL with distilled water. Keep at 4°C, and use within after 1 wk. (Larger 
quantities can be aliquoted and stored at -20 C C until use.) 

6. 5% Polyvinyl alcohol (PVA): Dissolve 5 g of PVA in 100 mL of PBS. Store at 4°C. 

7. 3-Aminopropyltriethoxysilane (APES)-coated slides: Place slides in racks and 
wash overnight in 10% decon-90. Wash in running hot water for 2 h. Dry over- 
night in oven at 50°C. Immerse slides in 2% solution of APES diluted in 100% 
alcohol for 10 min. Drain and wash in 100% alcohol for 10 min. Drain and wash 
in distilled water for 10 min. Repeat distilled water wash. Dry overnight in oven 
at 50°C. Store in clean boxes. 

8. ImmunoPure peroxidase suppressor (Pierce and Warriner, Chester). 

9. 3,3'-Diaminobenzidine (DAB) (Vector Laboratories). 

10. Hematoxylin: Gills hematoxylin (Sigma) dilute 1 :50 in distilled water before use. 

11. Diff-Quik stain (Baxter-Dade AG). 

12. Nikon 800 microscope equipped with a Basler digital camera and Lucia G image 
analysis software. 

3. Methods 

3. 1. Choice of Bone Samples 

Most of our studies have investigated protein expression in portions of 
human bone samples obtained from the iliac crest using a 8-mm trephine 
biopsy. Because adult bone is hard, brittle, and friable, bone samples are 
generally decalcified and wax embedded to obtain sections with the best mor- 
phology. Bone from younger patients is less brittle and therefore it is pos- 
sible to use frozen sections, which with careful cutting can provide good 
morphology. Some antibodies will work only on frozen sections. Whatever 
method is used it is important to know the site from which the bone sample 
was obtained and, for comparative studies between patients, to use bone from 
the same anatomical site. The age of the donor and any prescribed drug regi- 
men may also influence protein expression, and this should be considered 
when interpreting results. 

3.2. Preparing Wax-Embedded Sections 

1 . Place the biopsy on ice immediately after removal. 

17/Bord/237-248/F1 239 1 2/26/03, 10:47 AM 

I ■ I I I 


240 Bord 

2. If desired, divide biopsy into two halves to allow other types of analysis (see Fig. 
3 and Note 1). 

3. Fix the biopsy by placing in NBF overnight at 4°C (see Note 2). 

4. Remove the biopsy from NBF and wash in cold PBS for 2 h. 

5. Decalcify the bone sample by incubating in 14.5% buffered EDTA for 3 wk, 
replacing the EDTA every week, (see Note 3). 

6. Embed the biopsy in paraffin wax (see Note 4). 

3.3. Preparing Frozen Sections 

We usually prepare frozen sections in samples of neonatal ribs collected at 
postmortem from infants born at full term (30-40 wk). 

1 . Following removal, place the ribs immediately on ice. Trim the bone to an appro- 
priate size and dip the sample in chilled PVA for 2 min. 

2. Half fill a small plastic tray with chilled PVA and place the sample in the tray in 
the correct orientation for cryosectioning. 

3. Carefully lower the tray into liquid nitrogen, ensuring that the PVA solidifies 
before it comes into contact with the liquid nitrogen (see Note 5). 

4. Place the frozen sample in a 40-mL capped universal tube and store at -80°C. 

5. When required for cryosectioning, transfer the sample from -80°C to dry ice and 
carefully remove the peelaway plastic tray. 

6. Trim away excess PVA and mount on a suitable chuck for the cryostat (usually 
brass, 22-mm diameter). 

7. Prepare a chilling bath containing a 12-mm deep cold slurry of alcohol in chips 
of dry ice. Stand the chuck in this bath and pipet a thin layer of PVA onto the top 
of the cold chuck. 

8. Immediately place the bone sample (with the side to be sectioned uppermost) 
onto the PVA-covered surface of the chuck. More PVA can be added to ensure a 
tight bond between the PVA block containing the bone sample and the chuck. 

9. Return the block and chuck to dry ice. 

3.4. Sectioning Wax-Embedded Bone Samples 

1. Cut 7-|im thick sections on a base sledge microtome, float out on water (45°C) 
and collect on APES-coated slides. 

2. Dry the slides overnight at 45 °C and store in dust- free boxes. 

3. Immediately before performing immunohistochemistry, immerse the slides in 
xylene for 12 min. Replace with fresh xylene and incubate for another 12 min to 
dewax the sections. 

4. Transfer the slides to 100% ethanol and incubate for 3 min. 

5. Transfer the slides to 70% ethanol and incubate for 3 min. 

6. Transfer the slides to 50% ethanol and incubate for 3 min. 

7. Transfer the slides to PBS. 

17/Bord/237-248/F1 240 I 2/26/03, 10:47 AM 

I ■ I I I 


Protein Immunolocalization in Bone 24 1 

3.5. Cryosectioning Bone Samples 

The sections should be cut on a good quality cryostat equipped with a slow 
drive, high-torque motor, and automatic speed control at a cabinet temperature 
of-30°C (see Note 6). 

1. Cut 9-u.m sections and transfer onto APES-coated glass slides by placing the slide 
over the section on the knife blade and applying slight thumb pressure to the back 
of the slide for about 5 s. This encourages the section to adhere well to the slide. 

2. Air dry the slides for 10-20 min at room temperature. 

3. Place the slides in a staining trough and fix by incubating in 4% paraformalde- 
hyde for 30 min at room temperature. 

4. Wash three times in PBS for 5 min per wash. 

3.6. Immunohistochemistry Using an Indirect Immunoperoxidase 

The following procedure is suitable for both wax-embedded and frozen sec- 
tions. Several factors influence the choice of antibody and even once an anti- 
body is chosen, it is essential to establish the specificity of the immunoreaction 
to determine which control antibodies and which control tissues should be used 
to validate the procedure (see Note 7). 

1. Draw around the sections with an ImmEdge™ pen (Vector) to provide a fluid 
barrier and thus minimize the amount of reagents needed. 

2. Block endogenous peroxidase and nonspecific binding by incubation with ImmunoPure 
peroxidase suppressor (see Note 8). Carefully apply to the sections (about 100 uL/ 
section), taking care not to break over the hydrophobic ImmEdge™ pen barrier. 

3. Place the slides in a humid chamber for 30 min to prevent "drying out." 

4. Transfer slides to a staining trough and wash in PBS for 5 min with gentle shak- 
ing. Discard PBS by tipping the trough carefully so as not to dislodge sections 
from slides. Repeat the 5-min PBS wash two more times. 

5. Incubate sections with 10-20% serum in PBS for 30 min in a humid chamber to 
block nonspecific binding (see Note 9). 

6. Pour off the serum-PBS from the section, removing as much as possible. 

7. Dilute the primary antibody to a working concentration in 0.1% BSA and add 
about 100 uL/section (see Note 10). 

8. Incubate the slides in a humid chamber for either 90 min at room temperature or 
overnight at 4°C. 

9. Remove the slides from chamber to a staining trough and wash three times in 
PBS for 5 min per wash to remove unbound antibody. 

10. Remove the slides from PBS and apply the second antibody (3.5 ug/mL diluted 
in 0.1% BSA (see Note 11). 

11. Incubate the slides in a humid chamber for 30 min at room temperature. 

12. Prepare the ABC solution according to manufacturer's instructions and allow to 
stand for 30 min before use. 

17/Bord/237-248/F1 241 I 2/26/03, 10:47 AM 

I ■ I I I 


242 Bord 

13. Wash the sections in three changes of PBS (5 min each, with gently shaking) to 
remove the unbound secondary antibody and drain slides. 

14. Add ABC solution to the sections and incubate for 30 min at room temperature in 
a humid chamber. 

15. Wash the sections in three changes of PBS (5 min each, with gently shaking) and 
drain slides. 

16. Prepare the chromogenic substrate (DAB) according to the manufacturer's in- 
structions (see Note 12). 

17. Apply DAB to the section at timed intervals and then monitor microscopically 
for color development. Substrate action can be stopped by washing in distilled 
water and should be done at the same timed intervals as the application of the 

18. Wash the slides in distilled water. 

19. If required to detect nuclei and show general morphology, sections can be lightly 
counterstained with hematoxylin. Allow to blue in tap water for 10 min. Rinse in 
distilled water. 

20. Dry sections at room temperature, add Vectamount (Vector Laboratories), and 

21. Analyze using brightfield microscopy. 

3.7. Histological Staining 

It is important to save one or two sections cut at the start and end of the 
tissue block for histological examination as this provides morphological detail 
upon which to interpret the immunolocalization findings. These sections can 
be conveniently stained using Diff-Quick using the following protocol: 

1. Dip section in Diff-Quick solution A for 30 s. 

2. Drain and transfer to Diff-Quick solution B for 1 min. 

3. Drain and transfer to solution C for 1 min. 

4. Drain and rinse in distilled water until water clears. 

5. Allow to air-dry and mount with Vectormount. 

3.8. Quantitation of Findings 

1 . Immunostained sections should be examined under low-power microscopy, com- 
pared to the Diff-Quick stained sections, and areas of positive and negative stain- 
ing identified. 

2. The sections should next be examined under high power, noting the type and 
distribution of staining (nuclear, cytoplasmic, matrix, etc.). 

3. The results of the experiment can be analyzed and reported manually using a 
scoring system to denote the number of cells staining positively and/or the ob- 
served intensity of staining (e.g., +, ++, +++). Whenever possible, scoring should 
be carried out "blinded" to avoid reporter bias. 

4. As an alternative to manual analysis, the extent and intensity of staining can be 
analyzed semi-quantitatively by image analysis (see Note 13). 

17/Bord/237-248/F1 242 I 2/26/03, 10:47 AM 

I ■ I I I 

Protein Immunolocalization in Bone 



Fig. 2. Black-and-white photomicrograph demonstrating immunolocalization of 
ER-a by an indirect immunoperoxidase method with sites of antigenicity visualized 
with DAB (dark gray color). Positively stained osteoblasts (arrows), often adjacent to 
nonmineralized osteoid (*), are apparent on bone surfaces. Newly incorporated osteo- 
cytes (open arrows) show ER immunoreactivity. Marrow space cells (MS) surround 
the cancellous bone (bone). 

3.9 Practical Examples of Immunostaining in Bone Sections 

We have successfully used the aforementioned procedures to immuno- 
localize estrogen receptors (ERs) in cryosections of human neonatal bone (2) 
and transforming growth factor-|3 (TGF-|3) in sections from wax-embedded 
samples of adult human iliac crest bone (3). The former study investigated the 
differential protein expression of ERa and ER(3 in developing human rib (Fig. 
2). Intense ER expression was observed in osteoblasts and osteocytes in corti- 
cal bone. In contrast, ER-(3 was seen to be most highly expressed in osteoblasts 
and osteocytes in cancellous bone. The latter study provided quantitative data 
on the expression of TGF-|3, TGF-(3 receptors, platelet-derived growth factor 
(PDGF), and osteoclast activity in wax-embedded sections of adult human iliac 
crest bone samples. Bone sections from women treated with long-term high- 
dose estradiol were compared to those from women who had received no hor- 
mone replacement therapy. The results demonstrated that high-dose estrogen 
treatment is associated with increased TGF-(3, TGF-(3R, and PDGF synthesis 



2/26/03, 10:47 AM 

I ■ I I I 


244 Bord 

and decreased osteoclast activity, consistent with the hypothesis that these fac- 
tors may mediate the actions of estrogen in bone. These two studies were quite 
different in their approach, using the various techniques outlined in this chap- 
ter, and posed very different questions. The former study was an observational 
study conducted on frozen tissue fixed in 4% paraformaldehyde after section- 
ing. The latter study utilized NBF-fixed wax-embedded tissue with quantita- 
tive results obtained by image analysis measurements. However, both studies 
used the same immunoperoxidase localization techniques, demonstrating the 
versatility and reliability of this method. 

4. Notes 

1. Dividing biopsies for multiple uses: The biopsy can be halved longitudinally by 
placing the biopsy in a cylindrical shaped jig fitted with a thumbscrew to hold the 
biopsy in place. A slot on each side allows a fine modelmaker's saw to cut the 
through the biopsy. The thumbscrew is released and the 2 pieces of biopsy care- 
fully removed. One half can then be used for histomorphometry and the other 
half for histochemistry or RNA extraction (Fig. 3). 

2. Choice of fixative and antigen retrieval techniques: A detailed description of the 
different fixatives that can be used and their modes of action can be found in 
textbooks of histochemistry and histopathology (4,5). It is important to realize 
that the crosslinking action of fixatives may sometimes prevent access of anti- 
bodies to its antigen. There are many procedures available to unmask antigens 
(often referred to as "antigen retrieval"). Proteolytic digestion with enzymes such 
as trypsin, pronase, and pepsin is widely reported. We have successfully used 
Protease XXIV (dissolve 12.5 mg of Protease XXIV [Sigma] in 100 mL of 
PBS. Incubate sections for 30 min at 37°C prior to the ImmunoPure blocking 
step). Other methods of antigen retrieval use pressure, heat, and microwave 
irradiation with reported good results. However, it should be borne in mind that 
all these procedures may have deleterious effects on the adhesion of the section 
to the slide. 

3. Decalcification: After 3 wk in EDTA most bone biopsy samples are fully decal- 
cified; however, larger bone samples may require a longer time in EDTA. If 
desired, the extent of decalcification can be assessed radiographically. 

4. Wax embedding: Wax-embedding procedures require specialist equipment, not 
normally available in research laboratories. Our samples are embedded by the 
histopathology department, a service provided by many hospitals. Wax blocks 
should be stored at 4°C. 

5. Techniques for freezing samples: There are many methods of freezing tissue 
samples, critically reviewed by Chayen and Bitensky (6). It is important to choose 
a method that avoids the introduction of ice crystals, prevents cracking during 
freezing, and avoids cell damage, which leads to denaturing of the cytoplasm and 
loss of membrane integrity. Embedding bone samples in PVA prevents the tissue 
desiccation associated with low-temperature storage. 

17/Bord/237-248/F1 244 I 2/26/03, 10:47 AM 

I ■ I I I 

Protein Immunolocalization in Bone 


iliac crest bone biopsy 

— n t 

1 cm 

resin embedded for 

paraffin wax-embedded 

for immunolocalisation studies 

frozen for cryosections 
or molecular studies 

Fig. 3. Iliac crest bone biopsies can be cut longitudinally in a jig to provide two 
halves: one for resin-embedding for histomorphometry and one for either wax-embed- 
ding or freezing. 


6. Cryosectioning techniques: We use a Bright Cryostat equipped with a highly 
polished tungsten carbide knife. The angle of the knife is important, and we gen- 
erally set this to 19°C. The chuck and bone sample should be allowed to equili- 
brate to the temperature of the cryocabinet prior to cutting. Orient the block so 
that the smallest edge hits the knife first. Trim away carefully until full sections 
are obtained. Move the knife to expose the block to an unused cutting edge. 

7. Choice of antibodies and control antibodies: Both monoclonal and polyclonal 
antibodies can be used successfully for immunostaining bone sections, provided 
that they bind specifically to the antigen, or epitope of interest. Monoclonal anti- 
bodies (MAbs) usually show high specificity and low background staining. They 
work well on frozen tissue fixed in acetone or paraformaldehyde but are not so 
reactive in tissue that have been subjected to long fixation times or decalcifica- 
tion. Antigen retrieval in such tissue may sometimes unmask the epitopes (see 
Note 2). Polyclonal antibodies usually give a strong staining signal as the antis- 
era contain antibodies that react with different epitopes on the same antigen. Often 
they are associated with higher background staining than MAbs due to unknown 
reactivity of irrelevant antibodies from the immunized animal. Polyclonals usu- 
ally work well in wax-embedded tissue despite some denaturation having taken 
place. The choice of control antibodies is dependent on the type of primary anti- 
body used. With MAb it is possible to use another MAb directed against an anti- 
gen known to be absent from the tissue under investigation. We have successfully 
used a urease MAb as a control in our bone sections. With polyclonal antibodies 
it is usual to employ preimmune serum from an animal of the same species from 
which the primary antibody was generated. Care should be taken to ensure that 



2/26/03, 10:47 AM 

I ■ I I I 


246 Bord 

the control IgG concentration matches that used for the primary antibody. Fur- 
ther specificity checks can be made by using sections of tissue known to express 
and tissue known not to express the antigen under investigation. When carrying 
out these evaluations care should be taken to treat all tissue in the same manner, 
to minimize the chance of false negatives or positives. Some commercially avail- 
able antibodies are now available with the appropriate immunizing peptide. Pri- 
mary antibody is incubated with a 10-fold excess of immunizing peptide at room 
temperature for several hours prior to addition to the section. The antibody and 
immunizing peptide will complex, thus preventing binding to the antigen in the 
section, and give a negative signal. A positive signal would indicate that the pri- 
mary antibody is binding nonspecifically. 

8. Inhibiting endogenous peroxidase activity: Peroxidase enzymes present within 
the tissue can give false positive results. Hydrogen peroxide mixed in methanol 
or ethanol provides an efficient blocking agent but is very unstable at room tem- 
perature. We have therefore used Immunopure (store at 4°C) which is ready to 
use and stable at room temperature for the duration of this part of the experiment. 

9. Blocking nonspecific binding: The serum used for blocking prior to addition of 
the primary antibody should be from the same species from which the second 
antibody has been generated. The optimal concentration can be determined by 
trial and error. 

10. Optimization of primary antibody concentration: Most commercial antibodies 
are supplied with a data sheet that gives the recommended concentrations. Some- 
times these work well without further adjustment, but more often than not, the 
antibody concentration needs to be optimized for a particular tissue. This can be 
achieved by carrying out the procedure with a range of antibody concentrations. 
Staining should decrease with decreasing concentrations of applied antibody. If 
inappropriate staining occurs or the signal does not decrease at the lowest con- 
centrations of primary antibody it is essential to determine the cause. To ascer- 
tain the cause of nonspecific staining the procedure should be repeated with 
omission of specific components of the staining assay. To achieve this, four sec- 
tions should be set up; to tissue section 1 add only substrate. To section 2 add 
ABC and substrate. To section 3 add second antibody, ABC, and substrate. To 
section 4 add primary antibody, second antibody, ABC, and substrate. Only sec- 
tion 4 should show staining. If staining occurs on the other slides it is necessary 
to add further blocking steps to eliminate this inappropriate staining at the prob- 
lem stage. If sections 1-3 are negative but section 4 shows too much signal, 
decrease the primary antibody concentration. It may also be necessary to increase 
the concentration of the blocking serum if decreasing the primary antibody does 
not give the desired results. Too little or weak staining may indicate loss of anti- 
genicity of the primary antibody. This can be checked by repeating the procedure 
on a tissue known to express the antigen. Other problems should also be consid- 
ered. The activity of the substrate can be confirmed by mixing a few of the 
enzyme (ABC) with some substrate. An immediate color change should occur. 
Too high a concentration of second antibody can diminish signal. A different 

17/Bord/237-248/F1 246 1 2/26/03, 10:47 AM 

I ■ I I I 


Protein Immunolocalization in Bone 247 

fixative can be tried (e.g., acetone), as certain fixatives may mask epitopes, 
necessitating the use of antigen retrieval techniques. 

11. Choice of second antibody: The second antibody should be raised against the 
species from which the primary antibody is generated. Data sheets will indicate 
the concentration at which the antibody should be used. If nonspecific staining is 
a problem, 1% serum from the same species as the tissue being studied can be 
added to the second antibody as an additional blocking agent. 

12. Choice of chromogen: Several chromogens are available for immunoperoxidase 
procedures but DAB is the one we use as the brown color reaction provides a 
good contrast to the gray color of bone sections. Also, slides may be permanently 
mounted using this substrate. Caution: Care must be taken when handling DAB 
as it is a known carcinogen, but the hazards of using the compound are reduced 
when used in the kit supplied from Vector Laboratories. 

13. Image analysis software packages are becoming more widely available and have 
the advantage of removing the subjective nature of manual scoring. However, 
care should be taken to ensure that all color thresholds are determined and reflect 
accurately the positive signal. These thresholds should be used for all measure- 
ments within each experiment. All sections should be measured in a similar man- 
ner. Comparisons between experiments and between antibodies are to be avoided. 
Antibodies have different binding affinities, and therefore it would be inaccurate 
to say that one antigen is more highly expressed than another without other 
extensive studies. When making comparisons between different samples it is 
important that these samples should be analyzed in a single run with the same 
reagents. Particular care must be taken to ensure that all samples within the 
experiment receive the same incubation and development times. 


1. Hsu, S. M., Rainem L., and Fangerm H. (1981) A comparative study of the per- 
oxidase-antiperoxidase method and an avidin-biotin complex method for study- 
ing polypeptide hormones with radioimmunoassay antibodies. Am. J. Clin. Pathol. 
75, 734-738. 

2. Bord, S., Horner, A., Beavan, S. R., and Compston, J. E. (2001) Estrogen recep- 
tors a and (3 are differentially expressed in developing human bone. /. Clin. 
Endocrinol. Metab. 86, 2309-2314. 

3. Bord, S., Beavan, S., Ireland, D., Horner, A., and Compston, J. E. (2001) Mecha- 
nisms by which high-dose estrogen therapy produces anabolic skeletal effects in 
postmenopausal women: the role of locally produced growth factors. Bone 29, 

4. Hopwood, D. (1980) Fixation and fixatives, in Theory of Histological Techniques, 2nd 
ed. (Bancroft, J. D. and Stevens, J., eds.), Churchill Livingstone, London, pp. 20-40. 

5. Pearse, A. G. E., ed. (1968) Chemistry of fixation, in Histochemistry, Theoretical 
and Applied, vol. 1. Little, Brown, Boston, pp. 70-105. 

6. Chayen, J. and Bitensky, L., eds. (1991) Practical Histochemistry 2nd ed. John 
Wiley & Sons, London. 

17/Bord/237-248/F1 247 I 2/26/03, 10:47 AM 

I ■ I I I 




2/26/03, 10:47 AM 

I ■ I I I 



Detection of Noncollagenous Bone Proteins 
in Methylmethacrylate-Embedded Human 
Bone Sections 

Johannes P. T. M. van Leeuwen and Pieter Derkx 

1 . Introduction 

The organic matrix of bone is a well-organized network of proteins. The 
main constituent is type I collagen. The noncollagenous proteins (NCPs) com- 
prise about 10% of the total bone protein content. A variety of NCPs has been 
identified, including osteocalcin, osteopontin, osteonectin, bone sialoprotein, 
decorin, and biglycan. Of this group only osteocalcin and bone sialoprotein are 
specific for bone, whereas the other proteins are also present in other, 
noncalcifying, tissues. The knowledge of their role in and their effects on bone 
metabolism is limited. It is thought that these proteins play a role in the regula- 
tion of mineralization, in the attachment of osteoblasts and osteoclasts to the 
bone matrix, and/or attraction of cells to the bone matrix. In addition, a second 
group of proteins can be considered as noncollagenous bone matrix proteins, 
consisting of so-called growth factors "stored" in the bone matrix and may be 
released during bone resorption. Of these proteins, distinct cellular effects have 
been described, for example, regulation of growth and differentiation of cells. 
They probably do not play a direct role in the initiation of crystal formation of 
progress and termination of the mineralization process but may play a role in 
the coupling between resorption and formation. Examples of this second group 
are transforming growth factor beta (TGF6), insulin-like growth factors, and 
bone morphogenetic proteins. 

So far it appears that both groups of NCPs are involved in bone modeling 
during development and growth and in adult bone remodeling and fracture heal- 
ing. It is likely that for an optimal regulation of bone metabolism they act in 
concert. To achieve this, a coordinated synthesis and localization in the bone 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


18/vanl_eeuwen/249-258/F1 249 I 2/26/03, 10:47 AM 

I ■ I I I 


250 van Leeuwen and Derkx 

matrix would be essential. To understand the role of the NCPs in bone metabo- 
lism and bone diseases it is important to assess their precise localization and 
quantity and preferably to monitor changes in localization and quantity in aging 
and disease states. A few studies have examined the localization of 
noncollagenous bone matrix proteins in animals and in adult human bone. 

Immunolocalization studies in combination with standard bone histomorph- 
ometry (1) will allow the study of the relationship between NCPs and active osteo- 
clasts/osteoblasts, mineralization front, osteoid seams, and so forth. In addition, 
these combined analyses will provide data on the localization of these proteins in 
patients with metabolic bone diseases and may add to the knowledge on the func- 
tion of these proteins. The protocol was set up to be able to perform immunohis- 
tochemical analyses to localize and quantify NCPs in adult human bone, in 
combination with bone histomorphometry. For this the focus was to set up a tech- 
nique that can also be applied on bone specimens previously embedded in plastic 
(methylmethacrylate [MMA]) and used for bone histomorphometry. 

2. Materials 

2. 1. Plastic Embedding (see Note 1) 

1. MMA monomer (Merck). 

2. 0.4 g of Lucidol CH 50-L (a catalyst of polymerization active at -17°C; AKZO 
Chemicals, The Netherlands). 

3. Plastoid N or dibutylphthalate. 

4. Benzoylperoxide. 

5. Plastoid N or dibutylphthalate. 

6. W,W-dimethyl toluidine (Merck, accelerator of polymerizaton; see Note 2). 

7. Technovit (Kulzer, Germany) (powder A and solution B: mix A:B = 2:1). 

8. Thymol crystals (BDH). 

2.2. Preparation of MMA Mixture for-17°C Embedding 

Mix the following components in this order: 

1. 40 mL of stabilized MMA monomer. 

2. 0.4 g of Lucidol CH 50-L. 

3. 10 mL of Plastoid N or dibutylphthalate. 

2.3. Preparation of MMA Mixture for 37° C Embedding 

Mix the following components in this order: 

1. 40 mL of stabilized MMA monomer (Merck). 

2. 1.4 g of dried benzoyl peroxide (initiator of polymerization active at 37°C). 
Remove explosive-inhibiting moistness by placing a few grams of powder over- 
night at 37°C. 

3. 10 mL of Plastoid N or dibutylphthalate. 

18/vanl_eeuwen/249-258/F1 250 1 2/26/03, 10:47 AM 

I ■ I I I 


Noncollagenous Bone Proteins 251 

2.4. Technovit Mixture 

Mix powder A and solution B: A:B = 2:1. 

2.5. Immunostaining 

1. Primary antibodies of choice (see Note 3). 

2. 0.1% TPBS: phosphate buffered saline (PBS), pH 7.4, with 0.1% Triton X-100. 

3. 10% Normal goat serum (DAKO PATTS, Denmark) in 0.02% TPBS containing 
1.5% bovine serum albumin (BSA). 

4. Linker and label of the Biogenex Stravigen Multilink HRP kit (San Ramon, CA, 
USA) (see Note 4). 

5. 3-Amino-9-ethylcarbazole (AEC) (Sigma Chemical, St. Louis, MO, USA). 

6. 0.01% H 2 2 in 0.2 M of sodium acetate buffer pH 4.6. 

7. 0.2 M of sodium acetate buffer, pH 4.6: 13.3 g of sodium acetate trihydrate extra 
pure (Merck) in 1000 mL of distilled water; add 5.8 mL of glacial acetic acid, 5 mL 
of Tween 20, and stir for a few min. 

8. 1% AEC stock solution: 1 g of 3-amino-9-ethylcarbazole (Caution: carcinogen) 
in 100 mL of AA-dimethyl formamide (use a glass container). 

9. 1% AEC working solution (prepare fresh): Add 5 mL of AEC working solution 
to 95 mL of 0.2 M of sodium acetate buffer. Filter (wide-pore filter) and add 100 jxL 
of 30% H 2 2 just before use. 

10. Methyl green staining solution: Dissolve 1.65 gof methyl green powder (Nustain) 
in 250 mL of hot distilled water. Cool at room temperature. Remove purple color 
(= methyl violet component) by eight washes in 500 mL of chloroform. 

1 1 . Thionin stock solution: Heat 0.275 g of Thionin-M3 (C.I. 52000 Waldeck GmbH 
& Co. Division CHROMA Germany, in 100 mL of distilled 
water at 60°C for 20 min. The powder should be completely dissolved. Cool. The 
solution can be kept at 4°C for 4 wk. 

12. Citrate buffer: Add 1.25 gof anhydrous disodium hydrogen phosphate in lOOmL 
of distilled water to 0.63 g of citric acid in 300 mL of distilled water until pH 5.8 
is reached. A few drops of 2% sodium azide (toxic) should be added it stabilize 
the solution. This solution can be kept for 2-3 mo at 4°C). 

13. Thionin working solution: Mix one part thionin stock solution with five parts of cit- 
rate buffer. Heat to 37°C. The pH should be 5.8. Solution can be used for only 1 d. 

14. Light green staining solution: Mix 0.3 g of light green with 150 mL of distilled water, 
add 0.3 mL of glacial acetic acid, pH 2.9. Can be kept for 3 wk at room temperature. 

3. Methods 

3. 1. Bone Biopsy Preparation 

1. Bone material can be obtained from various origins. This procedure originally 
was set up to be applied on human bone but it is anticipated that it is applicable to 
bone tissue of other species. Osteoporotic iliac crest biopsies, a knee bone frag- 
ments (stress fracture), bone biopsies from a distal radius, and normal iliac crest 
biopsies from adult individuals were used. 

18/vanl_eeuwen/249-258/F1 251 1 2/26/03, 10:47 AM 

I ■ I I I 


252 van Leeuwen and Derkx 

3.2. Fixation 

1. Fix biopsies at room temperature for 24 h under vacuum (-0.85 bar) in either one 
of the following fixatives (see Note 5): 

a. 80% Alcohol. 

b. 4% Phosphate-buffered formaldehyde, pH 6.9. 

c. Burkhardt fixative, pH 7.4. 

3.3. Dehydration 

Dehydrate and impregnate fixed bone specimens either manually or using 
an automatic tissue processor according to the following scheme: 

1. 80% Ethanol, 1 h. 

2. Two times 96% ethanol, 1 h each. 

3. Two times 100% ethanol, 1 h each. 

4. 100% Ethanol, 10 h. 

5. Two times MMA (monomer), 1 h each. 

6. MMA (monomer), 7 h (at least). 

3.4. Embedding in MMA at-ITC 

Here we describe embedding done at -17°C, our current method, but we 
have in the past performed embedding at 37°C successfully (as described in 
Subheading 3.4.) (see Note 6). 

1. Add specimen and 15 mL of MMA mixture to a 20-mL glass vial (Packard) and 
place it under vacuum (-0.85 bar) for 1 h. 

2. Replace the MMA mixture with 15 mL of fresh MMA mixture and place again 
under vacuum (-0.85 bar) for 1 h. 

3. Cap vial, place it in a box with 70% ethanol and place the box at -17°C. 

4. After 1 h take the vial and add 150 uL of W,W-dimethyl toluidine (accelerator of 
polymerization) to the 15 mL of cold MMA mixture and shake gently, cap vial, 
and place it immediately in the 70% ethanol box at -17°C and leave overnight. 

5. The next day, first check whether polymerization is complete. If not, leave it 
longer at -17°C. 

6. When polymerization is complete, uncap the vial and discard the remaining liq- 
uid. To prevent breaking of the glass vial proceed carefully with the steps that 

7. Place capped vial in the 70% ethanol bath and place this in a refrigerator at 4°C 
for 1 h. 

8. Place it at room temperature for at least 1 h, followed by placing bath and vials at 
room temperature for at least 1 h. 

9. Uncap vial and cover sticky MMA upper layer by pouring a few milliliters of 
freshly prepared Technovit 3040 in the vial. This mixture will polymerize within 
15 min at room temperature. 

18/vanl_eeuwen/249-258/F1 252 I 2/26/03, 10:47 AM 

I ■ I I I 

Noncollagenous Bone Proteins 253 

3.5. Embedding in MMA at 3TC 

1. Add specimen and 15 mL of MMA mixture to a 20-mL glass vial (Packard) and 
place it under vacuum (-0.85 bar) for 1 h. 

2. Cap the vial and place it in a box with water and place this in an oven at 37°C for 
at least one night, until polymerization has been completed. 

3. Uncap the vial and cover sticky MMA upper layer by pouring a few milliliters of 
freshly prepared Technovit 3040 in the vial, this mixture will polymerize within 
15 min at room temperature. 

3.6. Sectioning 

1. Carefully remove embedded tissue from the glass vial by wrapping the cooled 
vial in a tissue and breaking the glass with a small hammer. 

2. Cut moistened sections on a Reichert-Jung Polycut S sliding microtome equipped 
with a Reichert-Jung tungsten carbide knife at an angle of 40°. 

3. Mount sections on gelatine-chromium-alum coated slides (see Subheading 3.8.) 
using a mixture of egg whites and glycerin (see Subheading 3.7.). 

4. For "immuno" slides: Stretch the sections with a 2:3 mixture of ethylene glycol 
monbutyl ether and 30% ethanol on a hot plate at 40°C (see Note 7). 

5. For "histomorphometrical" slides: Stretch the sections with 30% ethanol on a hot 
plate at 70-80°C or use the "immuno" slides preparation described in step 4 (see 

-^- Subheading 3.14.). 

3.7. Preparation of Egg White-Glycerin Solution 

1. Take a fresh egg, break it, and separate white of egg from the yolk. 

2. Put white of egg in a glass and measure volume (will be approx 15-25 mL). 

3. Mix egg white until it is stiff (manual mixer). 

4. Add an equal part of glycerin and mix again (± 30 sec). 

5. Add some Thymol crystals as antibacterial agent and stir again. 

6. Cover glass with parafilm and place solution at 4°C overnight. 

7. The next day, collect 15-30 mL of the yellow liquid into a small bottle (use a 
funnel) and store at 4°C. This solution keeps for several months. 

3.8. Mounting on Gelatin-Chromium-Alum-Coated Slides 

1 . Set heating plate to 40°C for immunoslides, to 70-80°C for histomorphometrical 
slides (see Subheading 3.6.). 

2. Place one drop of egg white-glycerin on a coated slide and rub out with a tissue. 

3. Dip section in 30% ethanol and put it down on the slide. 

4. Add some drops of 30% ethanol for immunoslides or a 2:3 mixture of ethylene 
glycol monobutyl ether and 30% ethanol for histomorphometrical slides: "swim- 
ming section" (see Subheading 3.6.). 

5. Put slide on heating plate, stretch with two dissection needles by rolling the needle 
from the middle of the section toward the left and right direction (removing air 
bubbles and flattening). It is very important to keep the slide moist! 

18/vanl_eeuwen/249-258/F1 253 1 2/26/03, 10:47 AM 

I ■ I I I 


van Leeuwen and Derkx 


6. Cut a strip of transparent cling film to the slide dimension (use the cheapest vari- 
ety, see Note 8) and put it on the slide. 

7. Pour off the excess ethanol carefully. 

8. Put the slide back on the heating plate. 

9. Flatten the section by rubbing off with filter paper using thumb or fingers. 

10. Place slides between two 5-mm pieces of wood. 

11. Clamp under light pressure for one night in a 45°C oven. Place one clamp in the 
center and or two clamps, one at either end. 

3.9. Removal of MM A (see Note 9) 

1 . Remove MMA from bone sections with a 1 : 1 mixture of xylene and chloroform 
for 30 min under constant stirring (important) in a setup as shown in Fig. 1. 

2. After 30 min place slides in xylene for 1 min. 

3. Wash slides three times in 100% ethanol, 20 sec each. 

4. Wash slides two times in 96% ethanol, 20 sec each. 

5. Wash slides two times in 70% ethanol, 20 sec each. 

6. Wash slides in 50% ethanol, 20 sec. 

7. Wash slides in distilled water, 20 sec. 

3.10. Decalcification (see Note 10) 

1. Sections are decalcified by incubation for 10 min in 1% acetic acid in the device 
shown in Fig. 1 under constant stirring. 

2. Rinse slides twice with distilled water. 

3.11. Immunochemistry 

The protocol for immunostaining given here works for a range of antibodies 
(2). All incubation steps (blocking, antibody incubation, and detection) should 
be carried out in a closed humidity chamber with the glass slides placed hori- 
zontally. The volumes used per section are 100-200 ^L. Capillary incubation 
systems (such as Techmate) in our hands result in lower staining intensity. 

1. Inhibit endogenous peroxidase activity by incubating the sections in a 10:1 mix- 
ture of methanol and 30% H 2 2 for 25 min at room temperature. 

2. Wash slides two times in distilled water. 

3. Rinse slides in 0.1% TPBS (see Note 11). 

4. Block with 10% normal goat serum in 0.02% TPBS containing 1.5% BSA (see 
Note 12). 

5. Remove excessive liquid by placing the edge of the glass slide on a tissue. 

6. Incubate with primary antibody in 0.02% TPBS containing 1.5% BSA overnight 
at 4°C. 

7. The next day, wash sections for 5 min with three changes of 0.1% TPBS. 

8. Incubate sections with 1:50 diluted Linker in 0.02% TPBS + 1.5% BSA for 
45 min. 

9. Wash three times with 0.1% TPBS, 5 min each. 

1 8/vanl_eeuwen/249-258/F1 


2/26/03, 10:47 AM 

I ■ I I I 

Noncollagenous Bone Proteins 


■ Lid 



■ Slides 

Magnetic stirrer 

- Lipshaw slide holder 

- Perforated 
metal "table" 

Fig. 1. Device for removal of MMA and decalcification. 


10. Incubate with 1:50 diluted label in 0.02% TPBS + 1.5% BSA for 45 min (Label = 
peroxidase-conjugated biotin streptavidin). 

11. Wash sections for 30 min with 0.2 mol/L of sodium acetate. 

12. Incubate with 0.05% AEC and 0.01% H 2 2 in 0.2 mol/L of sodium acetate 
trihydrate buffer, pH 4.6, for at least 15 min and check staining intensity under 
microscope (see Notes 13 and 14). 

13. Rinse sections with 0.2 mol/L of sodium acetate trihydrate buffer for 10 min. 

14. Rinse with running tap water followed by distilled water. 

15. Counterstain with methyl green, 15 sec. 

16. Rinse in distilled water, two times, 10 sec each. 

17. Remove excessive liquid by placing the edge of the glass slide on a tissue, for a 
few seconds. 

18. Mount in glycerin immediately. 

19. Seal glass edges with nail polish or hot wax. 

3.12. Modification of Staining Protocol for Quantitative Analyses 

When quantitative analysis is required use the staining protocol below (3): 

1. After step 14 in Subheading 3.11. continue as follows: 

2. Incubate slides in thionin for 2 min. 

3. Rinse shortly with distilled water. 

4. Dip the sections for 3 s in 0.2% light green solution. 

5. Rinse shortly with distilled water. 

6. Remove excessive liquid by placing the edge of the glass slide on a tissue, for a 
few seconds. 

1 8/vanl_eeuwen/249-258/F1 


2/26/03, 10:47 AM 

I ■ I I I 


256 van Leeuwen and Derkx 

7. Mount in glycerin. 

8. Seal glass edges with nail polish or hot wax. 

3. 13. Digital Quantitative Analyses 

After staining the sections can be digitized and quantitatively analyzed. The 
given protocol and equipment has been shown to be successful. However, other 
microscopes, cameras, and software platforms might also be applicable (see 
for example Chapter 24 by van 't Hof, this volume). 

1. Digital images of immunostained samples are captured by a three-chip color 
charge-coupled device (CCD) camera, 512 x 512 pixels (Sony) mounted on a 
Zeiss Axioplan microscope (Zeiss, Oberkochen) with a x20 objective. 

2. The KS-400 (version 1.2) digital image analysis system (Kontron) is used for the 
analysis of images of 30 trabecular and 20 cortical bone areas in each of four 
5-um sections cut at steps of 100 jxm in the same bone biopsy. 

3. For determining the area of the mineralized matrix (green) and NCPs (red/brown), 
thresholds are set on the green and red channel. The thresholds are determined 
interactively on the basis of three different images. Then, the determined thresh- 
old is used to analyze automatically all recorded images of all sections that are 
stained in the same staining session. 

4. The area of the mineralized matrix is calculated by counting pixels and eventu- 
ally expressed as u.m 2 . 

5. The same procedure is used to estimate immunostained areas. The ratio of the 
area of immunostained regions and the area of the mineralized matrix are calcu- 
lated in each microscopic field. Analysis with osteopontin and osteonectin dem- 
onstrates that the measurements are more or less stable when the number of 
microscopic fields exceeds 50 (2). 

3. 14. Histomorphometry 

For histomorphometry in combination with immunolocalization, parallel sec- 
tions are stained with Goldner (3), a modified von Kossa (4) or tartrate-resistant 
acid phosphatase. Details of these staining procedures can also be found in the 
chapter by Everts et al. and Chapter 24 by van 't Hof et al., this volume. 

4. Notes 

1. All resin mixtures should be prepared just before use. Care should be taken when 
handling plastic monomers, wear gloves and work in a fume hood as much as 
possible. Dispose of waste carefully and in compliance with local rules. 

2. Dimethyl toluidine can be used instead of dimethyl aniline. 

3. We have used the protocol described successfully with the following antibodies: 
Rabbit polyclonal antibodies LF-7 to osteopontin, LF-32 to osteocalcin, 
BON-1 to osteonectin, LF-83 to bone sialoprotein, LF-51 to biglycan, and 
LF-30 to decorin. These were a generous gift of Dr. L.W. Fisher, National 

18/vanl_eeuwen/249-258/F1 256 1 2/26/03, 10:47 AM 

I ■ I I I 


Noncollagenous Bone Proteins 257 

Institutes of Health, Bethesda, MD, USA (5). We have also used rabbit 
polyclonal antibodies against human transforming growth factor |3 2 and |3 
from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and antibodies 
against human insulin-like growth factor-I from GroPep (Adelaide, Austra- 
lia). This protocol therefore has proven to be applicable for detection of a 
range of antigens in mineralized bone (2). 

4. Linker contains a mixture of biotinylated goat anti-immunoglobulin antibodies 
and reacts with mouse, rat, rabbit, and guinea pig immunoglobulins. Other spe- 
cies-specific detection systems will also be applicable. 

5. Satisfactory results can be obtained with all three types of fixatives. 

6. Immunohistochemical analysis of NCPs can be performed after both procedures 
of embedding. At present, we routinely embed at -17°C but we have previously 
performed successful immunostaining of NCPs in bone fixed in Burkhardt and 
embedded at 37°C, 30 yr ago. 

7. To obtain good immunosignal it is important that slides are stretched at maximal 40°C 
because at higher temperatures the efficiency of the antigen detection is reduced. 

8 . The cheaper types of cling film are best because they do not stick to the sections/slides. 

9. The major advantage of MMA is that after sectioning it can be completely 
removed. This provides optimal penetration of the antibodies and chemicals nec- 
essary for immunostaining and histochemical staining. In addition, there is no 
background staining of plastic remains. With nondeacrylated sections staining of 
the bone matrix was much weaker and no cellular staining is observed. 

10. Decalcification generally yields better immunostaining. 

11. For some antibodies it might be advantageous to pretreat the sections before 
blocking with 0.1% saponin (0.1 g of saponin in 100 mL of distilled water, pre- 
pare just before use) for 30 min at room temperature. After saponin treatment 
generally lower concentrations of antibody can be used. 

12. To avoid background staining preincubation with normal goat serum is essential. 

13. AEC staining requires mounting in glycerin, which does not prevent fading (within 1 
wk) of the AEC signal and of dissolution of the thionin-light green stain. However, 
this period is long enough to digitize and store the images. Water-based polymers 
cannot be used because they result in disrupted trabeculae due to shrinkage. 

14. 3,3'-Diaminobenzidine (DAB) can also be used but in general AEC gives a more 
intense staining, more contrast, and no background. 


1. Recker, R., ed. (1983) Bone histomorphometry , Techniques and Interpretation. 
CRC Press, Boca Raton, FL, pp. 17-22. 

2. Derkx, P., Nigg, A. L., Bosman, F. T., et al. (1998) Immunolocalization and quan- 
tification of noncollagenous bone matrix proteins in methylmethacrylated-embed- 
ded adult human bone in combination with histomorphometry. Bone 22, 367-373. 

3. Derkx, P. and Birkenhager-Frenkel, D.H. (1995) A thionin stain for visualizing 
bone cells, mineralizing fronts, and cement lines in undecalcified bone sections. 
Biotech. Histochem. 70, 70-74. 

18/vanl_eeuwen/249-258/F1 257 I 2/26/03, 10:47 AM 

I ■ I I I 


van Leeuwen and Derkx 

Birkenhager-Frenkel, D. H., Grieten, P., and van der Heul, R. O. (1977) A new 
staining method for bone, osteoid, and cells in undecalcified bone sections. Calcif. 
Tissue Int. 24, (Suppl. 4). 

Fisher, L. W., Stubbs, J. T. Ill, and Young, M. F. (1995) Antisera and cDNA 
probes to human and certain animal model bone matrix noncollagenous proteins. 
Acta Orthop. Scand. 66, 1-5. 


1 8/vanl_eeuwen/249-258/F1 


2/26/03, 10:47 AM 

I ■ I I I 



Fluorescence Imaging of Bone-Resorbing 
Osteoclasts by Confocal Microscopy 

Stephen A. Nesbitt and Michael A. Horton 

1 . Introduction 

Osteoclasts are large multinucleate bone cells with the capacity to degrade 
bone by the process of bone resorption and, thus, participate in the homeostasis 
of bone and calcium in the body (I). Imaging of osteoclasts can be performed 
by a variety of microscopy methods including light microscopy, electron 
microscopy, and atomic force microscopy (AFM) (2,3). These techniques, 
together with histochemical and immunocytochemical stains, enable the 
researcher to analyze the cellular structure and function of this complex cell 
type both in vivo within bone tissues and isolated in vitro in primary cell cul- 
tures (see Part II, Culture of Osteoclasts). 

Unlike electron microscopy and AFM, which require a high degree of tech- 
nical expertise, methods for light microscopy, including laser scanning confo- 
cal microscopy (LSCM), are generally quick, easy to apply, and allow the 
screening of large numbers of cells. Furthermore, the combination of fluores- 
cence immunostaining and LSCM has provided a powerful, "hands-on" sys- 
tem for cell imaging (Fig. 1) (4). Today, the simplification of the LSCM user 
interface has made these systems accessible to many researchers and, to date, 
more than 10,000 publications have cited the application of LSCM. 

An early application of fluorescence immunostaining and LSCM was that 
used by Lakakkorpi and colleagues in 1993 to assess cell adhesion sites in 
bone-resorbing osteoclasts (5). From these studies, it was apparent that confo- 
cal microscopy was a technique that showed several advantages for cellular 
imaging in the field of bone cell biology. First, these techniques enabled sec- 
tional views of the resorption sites to be collected without the need to cut the 
specimens physically by tissue sectioning on a microtome. This is particularly 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


19/Nesbitt/259-282/F1 259 1 2/26/03, 10:47 AM 

I ■ I I I 


Nesbitt and Horton 


Kr/Ar laser 

Lenses X10,x16 r X63 
Pinhole set for image 
resolution at 350 nm 

50-1 00% power 

/ X ^RSP580 ^VRSP66Q ^< Mirror 

/ ^TTD488/S68/647 ^^ ^^ ^^ 

triple stained with 



PMT detector 1 



PMT detector 2 

PMT detector 3 

Motorised stage 

FITC - green 

TRITC -red 

Cy5 - blue 


Digitised image 


Display and analysis 

Fig. 1. Imaging of fluorescently labeled specimens by laser scanning confocal mi- 
croscopy: a schematic diagram of protocols. A krypton-argon laser provides a bright 
light source of known wavelengths that illuminates the specimen in focus. It scans the 
sample to reduce the time it is exposed to the light source, and so minimizes the 
"bleaching" of the fluorescence stains. The tripled-stained sample is sequentially ex- 
posed to the laser light at wavelengths 488, 568, and 647 nm for excitation of the 
FITC, TRITC, and CY5 fluorochromes, respectively. Pinholes situated at the laser 
light source and detectors ensure that "out-of-focus" fluorescence is omitted and only 
light from the focal plane is collected within an optical section of 350-nm thickness. 
Mirrors target the emitted light form the FITC, TRITC, and CY5 fluorochromes (emis- 
sion wavelengths 518, 576, and 670 nm, respectively) through a series of filters. The 
filtered light is collected by the detectors and converted into electrical signals and 
displayed as a digitized image. The specimen is mounted on a motorized stage and 
allows multiple optical sections to be sequentially collected at various depths through- 
out the specimen. The images can be viewed in both xy and zx planes and displayed in 
various formats: single fluorescence, intensity maps, merged for the analysis of 
colocalization of signals, or stacked to show a through focus view of the sample. 



2/26/03, 10:47 AM 

I ■ I I I 


Confocal Microscopy of Osteoclasts 26 1 

advantageous, as the procedures used to process bone samples prior to cutting 
(demineralization and tissue embedding methods) can destroy cell morphol- 
ogy and render protein epitopes inactive or inaccessible to a majority of anti- 
body probes. Furthermore, imaging of osteoclasts in bone by conventional 
fluorescence microscopy can be limited by the excessive autofluorescence 
given by the mineralized matrix in bone, and by blurring of the object owing to 
its inherent thickness. The confocal microscope allows for these background 
signals to be assessed and removed and can penetrate, in focus, deep into a 
specimen; thus, it can provide clear images of osteoclasts and their subcellular 

Herein, the application of fluorescence immunostaining and LSCM meth- 
ods are described for the imaging and functional analysis of bone-resorbing 
osteoclasts. An in vitro model of bone resorption of human osteoclasts isolated 
from human osteoclastoma (giant cell tumor of bone) tissue and seeded onto 
biotin-labeled dentine was used (6). The cells were probed by triple fluores- 
cence staining and visualized with LSCM. The osteoclasts were identified by 
the presence of characteristic rings or arcs of F-actin staining with phalloidin- 
tetramethylrhodamine isothiocyanate (TRITC) and the resorption sites were 
identified by a loss of the biotin label from the dentine surface with 
streptavidin-CY5. Primary antibody probes together with secondary-labeled 
anti-immunoglobulin fluorescein isothiocyanate (FITC) antibodies enabled 
proteins of interest to be examined for their localization within osteoclasts in 
the ruffled border, sealing zone, basolateral plasma membrane, within the cell 
body, and in the resorption pit. The resorption sites were imaged as cellular 
sections taken in xy (above view) and zx (lateral view) planes with single- and 
triple-color displays and reconstructed as views in two- and three-dimensional 
formats (7). The lysosomal associated membrane protein (LAMP) (8), the cell 
surface receptor a v (3 3 integrin (9), and the metalloproteinase MMP-9 (10) were 
used as example molecules to demonstrate the application of these osteoclast 
imaging techniques both in vitro in resorption cultures and also in vivo, in 
thick cryosections of osteoclastoma. 

The application of these imaging techniques to functional analysis is further 
discussed and include the use of FITC labels to "track" the exogenous addition 
of compounds within the resorption cultures and measurement of resorption 
using fluorescence intensity profiles to quantify the amounts of degraded 
matrix endocytosed by osteoclasts during the bone resorption process (6,11). 

Fig. 1. (continued) Additional computation is required to convert the image stack into 
a 3D reconstruction. Image analysis can show fluorescence profiles and quantify fluo- 
rescence signals for the whole specimen or for discrete sites within the specimen. 

19/Nesbitt/259-282/F1 261 1 2/26/03, 10:47 AM 

I ■ I I I 


262 Nesbitt and Horton 

2. Materials 

2. 1. Buffers and Solutions 

1 . Tissue culture media: Minimum essential medium (MEM), 10% fetal calf serum (FCS), 
2% glutamine, penicillin, and streptomycin (GPS) at 100 U/mL (Gibco/Invitrogen). 

2. Phosphate-buffered saline-Dulbecco A (PBS), pH 7.2 (Oxoid, UK). 

3. Fixation buffer: 3.5% Paraformaldehyde, 2% sucrose in PBS (see the chapter by 
Everts et al., this volume, for details). 

4. Wash solution: PBS, 5% normal calf serum (Gibco/Invitrogen), 0.05% sodium 
azide, pH 7.2. 

5. Freezing buffer: PBS, 2.3 M sucrose, 0.05% sodium azide. 

6. Biotin staining buffer: 230 mM NaCl, 50 mM NaHC0 3 , pH to 8.2 with 1 M HC1. 

7. Triton permeabilization buffer: 0.5% Triton X-100 (Pierce, UK), 20 mM N-2- 
hydroxyethylpiperazine-W-2-ethanesulfonic acid (HEPES), 300 mM sucrose, 50 mM 
NaCl, 3 mM MgCl 2 , 0.05% sodium azide, pH 7.0. 

8. Saponin permeabilization buffer: 0.1% Saponin (Calbiochem, UK) in piperazine- 
Nfl'-bis (2-ethanesulfonic acid) (PIPES) buffer, pH 6.8. 

2.2. Reagents for Immunostaining and Protein Labeling 

1. Primary antibodies of choice. 

2. Secondary antibodies, for example, rabbit anti-mouse, swine anti-rabbit, or swine 
anti-goat immunoglobulin reagents (Dakopatts, Denmark). 

3. Streptavidin-Cy5 (Amersham/Pharmacia, UK). 

4. Phalloidin-rhodamine (Molecular Probes, Leiden, The Netherlands). 

5. Antifade: PBS Citifluor (Canterbury University, UK). 

6. Fluoro-Link reactive dye, FluoroX No. A28000 (Amersham, UK). 

7. Centricon columns, 3-kDa cutoff (Millipore, UK). 

2.3. Cells and Tissues 

1. Elephant tusk (see Note 1). 

2. Giant cell tumor of bone (osteoclastoma). 

2.4. Equipment 

1. ISOMET low-speed diamond saw (Buehler, USA). 

2. Leica scanning laser confocal microscope, system TCS NT (Leica, Heidelberg, 
Germany) equipped with a multiline 750 mW Omnichrome krypton-argon laser 
(Chino, USA) and objective lenses xlO (NA 0.30), -16 (NA 0.50), and x63- 
water immersion (NA 1.20). 

3. Silicon Graphics 2 Workstation (SGI, Mountainview, CA, USA) and Imaris 
software for image analysis (Bitplane, Zurich, Switzerland). 

3. Methods 

3. 1. Preparation of Dentine Discs 

Dentine provides a suitable matrix substrate to study osteoclastic bone re- 
sorption in vitro (12-14) (see Note 1). 

19/Nesbitt/259-282/F1 262 I 2/26/03, 10:47 AM 

I ■ I I I 

Confocal Microscopy of Osteoclasts 263 

3.1.1. Cutting and Steriliation of Dentine 

1. Using an electric bandsaw, cut the elephant tusk across its length at 5-cm inter- 
vals to give a series of large dentine discs. Rotate the discs 90°C and saw into 
four to eight equal segments (see Note 2). 

2. Slice the segmented blocks of dentine into fine wafers (100 [im in thickness and 
a surface area 2-3 cm 2 ) with a water-cooled, ISOMET low-speed diamond saw 
(Buehler, USA); operate at speed no. 4 and weight the dentine block with 125 g 
to cut flat slices of dentine. 

3. Rinse the wafers in distilled water and allow to soak for 1 h to soften the dentine 
before "hole punching" the wafers into 6-mm discs. Collect 200 discs. 

4. Polish both sides of the dentine disks by rubbing onto 3MM paper, each for 10 s. 

5. Mark the discs with a no. 2 pencil on one side. This assists to assess the orienta- 
tion of the disc in future experiments; place the discs with the mark on the lower 
surface when manipulated or when in culture. 

6. Wash in 50 mL of distilled water for 30 min on a roller mixer to remove debris. 
Repeat the washing procedure twice. 

7. Sterilized the discs in 70% ethanol for 5 s, and rinse three times in sterile distilled 
water (see Note 3). 

8. In a laminar flow tissue culture hood, dispense the discs into a large Petri dish 
(keep sterile) and allow them to dry. After 2 h, flip the discs and allow thorough 

-m- drying for another 2 h before storing at room temperature (RT). This is a conve- 

nient stop point. Alternatively, to avoid drying the discs, leave out step 8 and 
continue on to the biotinylation methods described in Subheading 3.1.2. 

3. 1.2. Surface Biotinyation of Dentine Discs 

The surface mineral is partly stripped during dentine cutting and a propor- 
tion of the matrix proteins are exposed at the dentine surface. Biotin provides a 
suitable label for these exposed proteins including type I collagen, which com- 
prises >90% of the organic matrix in bone and dentine (6). Biotinylation prod- 
ucts are stable and do not affect osteoclast resorption. Herein, the dentine 
proteins are quickly and easily labeled with a protein biotinylation kit using the 
biotinylation reagent biotinamidocaproate iV-hydroxysuccinamide ester (see 
Note 4). 

The ECL protein biotinylation kit (RPN 2202, Amersham/Pharmacia, UK) 
was primarily designed for biochemical applications, but in addition, it has 
proved useful for the surface labeling of dentine discs. The labeling should be 
performed under sterile conditions. 

1. Place 200 dentine discs in a 30-mL universal container and rehydrate in 10 mL of 
distilled water on a roller mixer for 10 min. Discard the wash and repeat twice. 

2. Add 10 mL of 50 mM NaHC0 3 buffer pH 8-8.5 (made fresh, from the 20x stock 
in the kit) and wash for 10 min. Discard the buffer and repeat the wash once. 
Resuspend in 10 mL of buffer. 

19/Nesbitt/259-282/F1 263 1 2/26/03, 10:47 AM 

I ■ I I I 


264 Nesbitt and Horton 

3. Allow the biotinylation reagent to equalibrate at RT for 30 min. Add 50 \xL of the 
biotinylation reagent, and gently invert/mix to resuspend the discs/buffer/label. 
Cover the universal container with foil to shield from direct light. Place on a 
roller mixer for 1 h and every 10 min gently invert/mix and return to the roller 
mixer (see Note 5). 

4. Discard the buffer and transfer the discs to a 50-mL tube for thorough washing. 
Resuspend in 40 mL of distilled water and wash vigorously on roller mixer for 30 
min. Discard the wash and repeat the washing five times. 

5. Place the biotinylated discs into a large Petri dish and add 200 mL of PBS with 
100 U/mL of penicillin-streptomycin and leave to soak overnight in a humidified 
incubator with 5% C0 2 . 

6. Remove the soaking solution and dry the biotinylated discs as described in Sub- 
heading 3.1.1., step 8 and store at 4°C in a dessicated box {see Notes 6 and 7). 

3.2. Preparation of Osteoclasts from Giant Cell Tumours of Bone 

Osteoclastomas (giant cell tumor of bone) are rare bone tumors, but provide 
an invaluable source of mature human osteoclasts that are otherwise difficult 
to extract from the hard matrix of bone. These cells are not the neoplastic cell 
type and have a normal phenotype and function (15,16). The osteoclasts iso- 
lated from osteoclastoma are generally large (>50 txm) and this helps to image 
the bone-resorbing osteoclasts and their subcellular structures by LSCM. 

The following method allows large numbers of viable osteoclasts to be ex- 
tracted from osteoclastoma tissues by digestion of the collagenous extracellu- 
lar matrix (17). Typically, 2 x 10 6 osteoclasts can be obtained from 5 g of 
osteoclastoma tissue. The cells can be placed directly into tissue culture or 
frozen for long-term storage. 

1. Perform all procedures in sterile conditions. 

2. Take the osteoclastoma tissue immediately after surgical removal and place into 
MEM with 2% glutamine and penicillin-streptomycin at 100 U/mL (MEM-GPS) 
and leave on ice. 

3. Dissect out the fat, connective, and necrotic tissues. 

4. Place 5 g of osteoclastoma into 5 mL of ice-cold MEM-GPS and chop into a fine 
tissue suspension. 

5. Add 10 mL of type I collagenase (Sigma, UK) at 3 mg/mL in MEM-GPS, mix, 
and incubate at 37°C for 30 min. Further mix to resuspend the tissue digest every 
5 min. Prewarm the collagenase solution for 10 min at 37°C before use. 

6. Pass the digested tissue through a sieve with a 100-jxm mesh (Falcon, UK). Use 
the plunger from a 5-mL syringe and 30 mL of MEM-GPS at RT to mix and 
flush the osteoclastoma cells through the mesh. Discard the undigested tissue left 
in the sieve, (see Note 8). 

7. Centrifuge the osteoclastoma cells at 200g for 10 min at RT. 

8. Resuspend and wash the cell pellet in 50 mL of MEM-GPS with 10% FCS and 
repeat step 7. Finally, repeat steps 8 and 7 at 4°C and place the cell pellet on ice. 

19/Nesbitt/259-282/F1 264 I 2/26/03, 10:47 AM 

I ■ I I I 


Confocal Microscopy of Osteoclasts 265 

9. Resuspend the cell pellet in tissue culture media (MEM-GPS and 10% FCS) and 
proceed with the resorption cultures in Subheading 3.3. 

10. Alternatively, freeze the osteoclastoma cells at 2 x 10 4 osteoclasts per cryovial in 
1 mL of MEM with 40% FCS and 10% dimethyl sulfoxide (DMSO). After rate 
freezing to -70°C, store the cells in liquid nitrogen. 

11. To retrieve the cells from frozen stocks, reconstitute a vial of frozen cells with 1 mL 
of ice-cold MEM-GPS and repeat step 7 at 4°C followed by step 9. The osteo- 
clast viability is approx 50% after freezing the cells. 

3.3. Resorption Cultures 

In vitro models to study osteoclastic bone resorption were initially described 
by Boyde et al. and Chambers et al. in 1984 and Arnett and Dempster in 1986 
(12-14). Osteoclasts are cultured on the surfaces of bone and dentine and form 
resorption pits that are similar to resorption sites seen in vivo. The number and 
size of the pits provide a measure of resorption and these models can be used to 
screen compounds that regulate bone resorption (18-20). 

Herein, human osteoclasts isolated from osteoclastoma tissue are cultured 
on biotinylated dentine discs and are used to study cellular structure and func- 
tion in bone-resorbing osteoclasts by fluorescence immunostaining and LSCM 
(6) (see Note 9). 

1. Soak the biotinylated dentine discs in culture media (MEM-GPS and 10% FCS) 
for 30 min. 

2. Remove the media and place the discs onto a sheet of sterilized parafilm in a Petri 
dish. Excess media remaining on the dentine surface can be removed by touching 
the edge of the disc on a flat sheet of tissue paper, but do not allow the discs to 
dry out, as this will inhibit cell attachment. 

3. Resuspend the osteoclastoma cells in culture media. Between 1% and 5% of the 
osteoclastoma cells are osteoclasts. Apply 100-200 osteoclasts per disc. Use a 1-mL 
pastette (Jencons, UK) and place a droplet of cells onto each disc. The discs hold 
approx 40 yiL and the droplets are held in position by the hydrophobicity of the 

4. Transfer the cells into tissue culture incubator, providing a humidified atmosphere 
of5%C0 2 at37°Cfor2h. 

5. Rinse off nonadherent cells by shaking the discs in a bath of PBS at 37°C (see 
Note 10). 

6. Place the discs and attached cells, which includes an enriched population of 
osteoclasts, into resorption media (MEM-GPS, 10% FCS, and 10 mM HC1 to 
provide an operating pH of 6.9) and return to tissue culture for 22 h. Resorption is 
optimal under these conditions (see the chapter by Hoeberts and Arnett, this vol- 
ume, for discussion of optimal pH for resorption). Use a 24-well tissue culture 
plate and place one to four discs per well with 1 mL of media (see Note 11). 

7. To remove cytosolic components from the cells refer to Subheading 3.4. Alter- 
natively, proceed onto step 8. 

19/Nesbitt/259-282/F1 265 1 2/26/03, 10:47 AM 

I ■ I I I 


266 Nesbitt and Horton 

8. Quickly rinse the discs in MEM and immediately add to each well 1 mL of a 
50:50 mixture of MEM with paraformaldehyde fixation buffer (3.5% paraform- 
aldehyde, 2% sucrose in PBS) at 37°C and leave for 10 min. Alternatively, fix the 
discs in ice-cold methanol for 6 min on ice (21) (see Note 12). 

9. Rinse the discs three times in PBS and leave in wash (5% NCS, 0.05% sodium 
azide, and PBS) at 4°C. 

10. Proceed to the cell staining methods described in Subheading 3.5. Alternatively, 
soak the discs in freezing buffer (2.3 M sucrose in PBS with 0.05% sodium azide) 
at 4°C for 4 h and store at -70°C. To retrieve the frozen discs, thaw at RT and 
repeat step 9 (see Note 13). 

3.4. Removal ofCytosolic Proteins from Osteoclasts During Cell Cul- 

Some cell proteins are expressed in high amounts in the cytosol compared to 
their membrane-associated forms. Thus, to see the membrane associated pro- 
teins it may be necessary to remove the cytosolic forms. This can be achieved 
by a temporary permeabilization of the cell membrane with saponin during 
tissue culture, with the cytosol being released into the media before cell fixa- 
tion (21,22). Previously, these procedures have been used in osteoclast cul- 
tures to visualize membrane bound annexins by fluorescence immunostaining 

1. Place the saponin permeabilization buffer (0.1% saponin, 80 mM disodium 
PIPES, pH 6.8) in a C0 2 incubator for 1 h. 

2. Rinse the resorption culture discs in MEM at 37°C. 

3. Add the prewarmed/pregassed saponin to the cells and return to tissue culture for 
5 min. 

4. Repeat step 2 and fix the cells as described in Subheading 3.3., steps 8-10. 

3.5. Multicolor Labeling of Cells by Fluorescence Immunostaining 

A variety of fluorescent probes are available to immunostain cells and sub- 
cellular structures and these include probes for the plasma membrane, nucleus, 
Golgi, mitochondria, lysosomes, and the cell's cytoskeleton (Molecular Probes, 
Holland; Various combinations of these stains allow mul- 
tiple cell types, cell structures, and cell proteins to be examined together. 

3.5.1. Cell Permeabilization 

Cell permeabilization is required after cell fixation with paraformaldehyde 
to allow intracellular access of the antibodies and fluorescence stains. 

1. Select the number of cultured discs required for staining. Use a 24- well tissue 
culture plate and place a cultured disc into each well and add 1 mL of wash (5% 
NCS, 0.05% sodium azide, and PBS) and leave on ice. 

19/Nesbitt/259-282/F1 266 1 2/26/03, 10:47 AM 

I ■ I I I 


Confocal Microscopy of Osteoclasts 267 

2. Rinse the discs in PBS, and permeabilize the cells for 5 min in 1 mL of ice-cold 
Triton X-100 permeabilization buffer (20 mM HEPES, 300 mM sucrose, 50 mM 
NaCl, 3 mM MgCl 2 , 0.5% Triton X-100, 0.05% sodium azide, in PBS at pH 7.0). 
Alternatively, use 0.1% saponin permeabilization buffer (0.1% saponin, 80 mM 
disodium PIPES, pH 6.8) for 30 min on ice. 

3. Rinse the discs twice in PBS and soak in wash for 30 min (see Note 14). 

3.5.2. Triple Fluorescence Immunostaining 

Triple fluorescence immunostaining of the resorption cultures enables the 
resorption sites and resorbing osteoclasts to be identified and characterized 
within the mixed cell population of osteoclastoma cells. Resorbing osteoclasts 
are identified by their characteristic sealing zones and are stained for F-actin 
with a phalloidin-rhodamine conjugate. The biotinylated matrix proteins are 
probed with a streptavidin-CY5 conjugate and the resorption pits stain nega- 
tive due to the loss of the biotinylated matrix from the resorption sites. Addi- 
tional proteins are stained with specific primary antibodies and secondary 
labeled anti-immunoglobulin FITC antibodies. 

Perform all subsequent cell rinsing/washing on ice in a 24-well tissue plate 
and the staining at RT in a wet box. Do not let the cells dry out at any stage of 
the staining procedure, as this will increase sample autofluorescence and el- 
evate background signals in the LSCM analysis. 

1 . Take the permeabilized cells and drain off the excess wash from the surface of 
the disc by touching its edge onto tissue paper. 

2. Place the discs on a sheet of parafilm in a wet box. 

3. Add 100 |xL of primary antibody (e.g., a mouse monoclonal, or a rabbit or goat 
polyclonal) in wash and probe for 30 min. The concentrations of the primary 
antibodies will vary, but generally, 5-10 p,g/mL, or a 1:100 dilution of polyclonal 
sera is a good starting point. 

4. Repeat step 1 and rinse the discs twice in PBS and leave to soak in wash for 30 min. 

5. Repeat steps 1 and 2 and add 100 jxL of FITC-conjugated secondary polyclonal 
antibody at a 1:40 dilution in wash and stain for 30 min. 

6. Repeat step 4. 

7. Repeat step 1 and add 100 uE of streptavidin-CY5 at 10 |ig/mL in biotin staining 
buffer (230 mM NaCl, 50 mM NaHC0 3 , pH to 8.2) for 60 min. 

8. Repeat step 4. 

9. Repeat step 1 and add 100 jiL of phalloidin-rhodamine conjugate at 5 U/mL in 
wash and stain for 30 min. 

10. Repeat step 4. 

11. Stain additional discs for negative controls. Repeat steps 1-11, but omit staining 
with primary antibodies, steps 3 and 4. 

12. Mount the triple-stained samples in PBS-antifade reagents and view by fluores- 
cence microscopy and LSCM (see Note 15). 

19/Nesbitt/259-282/F1 267 I 2/26/03, 10:47 AM 

I ■ I I I 


268 Nesbitt and Horton 

3.6. Laser Scanning Con focal Microscopy 

A variety of confocal microscopes are available and include systems manu- 
factured by Leica, Bio-Rad, Zeiss, and Olympus. Herein, a Leica TCS NT con- 
focal microcope (Leica, Heidelberg, Germany) equipped with a multiline 
750-mW Omnichrome krypton-argon laser (Chino, USA) was used to image 
osteoclasts after triple-color immunostaining (6,7). Excitation of the 
fluorochomes FITC, TRITC, and rhodamine, and CY5 was performed at 488, 
568, and 647 nm, respectively, and their optical images were collected, 
digitised, and displayed in pseudocolors of green, red, and blue, respectively 
(refer to Fig. 1 for a schematic diagram of LSCM). 

3.6.1. Setting up the Confocal Microscope 

1. Go to the acquisition interface. 

2. Adjust the filters for FITC, TRITC, and CY5 and activate their sequential collection. 

3. Select the appropriate objective lens, xl6 or x63. A xl6 objective lens is suitable 
for low-power overviews of the resorption sites and a x63 enables high-power 
images of individual osteoclasts and resorption sites to be examined. 

4. Set the mode to xy to collect optical sections in the horizontal plane. 

5. Keep the pinhole setting at 1 for optimal image resolution (> 350 nm). 

6. Set the collection speed to medium, the zoom to x 1 , and image format to 5 12, the 
section number to 16, and the section accumulations to 2. 

7. Adjust the laser to 50% power and photomultiplier tube voltage thresholds (PMT) 
to between 400 and 700 mV for channels 1 (FITC), 2 (TRITC). and 3 (CY5). 

3.6.2. Image Collection 

1. View the samples on the standard fluorescence microscope with the FITC and 
TRITC filters and light from the mercury UV lamp. Use a xlO objective to screen 
the sample. 

2. First check the negative controls samples for nonspecific staining. 

3. Identify the resorbing osteoclasts by their F-actin rhodamine staining with the 
568-nm filter (Fig. 2A). 

4. Select a representative area of osteoclastic resorption and view with a x 16 objective. 

5. Check the FITC staining with the 488-nm filter. Note, the CY5 647-nm wave- 
length is not visible with light from the mercury UV lamp and can be viewed 
only with the laser light source. 

6. Scan the negative control sample with the confocal microscope and adjust the 
PMTs for each channel to gate out the background signals from nonspecific stain- 
ing and autofluorescence from the cells and the dentine discs. 

7. Repeat steps 3-5 on the stained positive samples and scan. Lower the PMTs if 
the fluorescence signals for the images are too high and adjust to obtain a distinct 
staining pattern. Crosstalk between the channels may be visible on the scanned 
images, but is removed when the scan is, subsequently, operated in a series mode 
(step 9) (see Note 16). 

19/Nesbitt/259-282/F1 268 1 2/26/03, 10:47 AM 


.Throug^ffocus (A to. C) 

in in the . 

■ Resorption path.- " 

A Above view C 

50 f/rr\ 

C At^pve viev% 

Fig. 2. A low-power through focus overview of resorbing osteoclasts, cultured on labeled dentine. Human osteoclasts were 
cultured on biotinylated dentine for 24 h. The cells were fixed and permeabilized before fluorescent immunostaining and confocal 
microscopy {see Subheadings 3.3-3.6.). The through focus images were compiled from 16 optical sections (each 350-nm in 
thickness) taken through the sites of resorption in the xy plane and are displayed in gray scale. The single fluorescent images in (A), 
show F-actin; in (B), lysosomal associated membrane proteins (LAMPS); and in (C), the biotinylated dentine proteins. Examples 
of sealing zones and LAMPS in resorbing osteoclasts, and their resorption pits are arrowed in (A), (B), and (C), respectively, and 
the broad arrow shows the directional path of resorption taken by an osteoclast. A colored overlay of these images is shown in Fig. 
5 A. (Reduced from original magnification, xl60.) 















I ■ I I I 


270 Nesbitt and Horton 

8. While scanning, adjust the z-position for the microscope stage and set the top 
position at the osteoclast's apex and the bottom position at the base of the resorp- 
tion pit. Minimize the time of the scanning to reduce photobleaching of the sample 
by the laser light. 

9. Start the series collection. The system automatically collects a series of 16 equi- 
distant, optical sections from the cell's apex to the bottom of the resorption pit 
for each fluorescence channel in a sequential manner. Two accumulations are 
taken for each optical slice and averaged to reduce noise in the image. 

10. Save the data. A typical file size is 12 MB for a stack of 16, triple-stained images. 

11. View the stack of images as single- and triple-stained images (Figs. 2-4, and 5 
and 6, respectively) in gallery and quarter-tile formats. Select and display the 
images as described in Subheading 3.6.3. 

12. Change the objective to x63 and view the sample with water immersion. 

13. Repeat steps 6-11 and collect high-power images of individual sites of osteo- 
clast resorption (Figs. 3 and 5). 

14. Repeat step 13 and adjust the laser to full power and collect 64 optical sections to 
obtain image data for three-dimensional analysis of the resorbing osteoclasts {see 
Subheading 3.6.3., step 11) (Fig. 6). Photobleaching of the scanned area will be 
extensive and this area is not suitable for further imaging. Select other areas of 
the sample for more imaging. 

15. Change the mode to zx (vertical) plane to observe the lateral view image. Repeat 
steps 6-11 and collect lateral views of the resorption sites (Figs. 3D and 5C). 
Imaging in the zx plane can cause "scan lines" from photobleaching. These "scan 
lines" will appear in subsequent imaging of the same cells in the xy plane, there- 
fore, complete the xy imaging before scanning in the zx plane. 

3.6.3. Image Display 

1. Go to the view interface. 

2. Display the single fluorescence images in gray scale (Figs. 2-4) and triple-stained 
images as pseudocolored overlays (Fig. 5). 

3. View the low-magnification images (xl60) in through focus for the entire image 
stack and to obtain an overview of the resorption sites (Figs. 2A-C and 5A). 

4. View the high-magnification images (x630) in each of the optical sections and 
analyze the osteoclast and its sub-celluar staining in detail in the xy plane (above 
view) (Figs. 3A-C and 5B,D) and zx plane (lateral view) (Figs. 3D, 5C). 

5. Note the localization and intensity of the stains in the resorption pit, along the 
resorption surface, in the resorbing osteoclasts at the cell's surface, ruffled bor- 
der, sealing zone, cytoplasmic skirt, and in the basolateral cell body. In addition, 
examine the staining patterns in the surrounding stroma and in the nonresorbing 
osteoclasts (those osteoclasts without sealing zones and resident on the intact 
surface of dentine). 

6. Identify colocalization of stains. For example, on merging the images green and 
red gives yellow to orange; green and blue gives pale blue and pale green; red 
and blue gives pink; and together green, red, and blue gives white (Fig. 5). 

19/Nesbitt/259-282/F1 270 1 2/26/03, 10:47 AM 

I ■ I I I 


Fragcrtents of degrade 
bone matrix within 
the osteoclast 



Cell top 


Above view 

~31 ! 

Resorption "" / 

path Ruffled '« ceil base 

Side view 

Fig. 3. High-power sectional views of a resorbing osteoclast cultured on labeled 
dentine. Human osteoclasts were cultured on biotinylated dentine for 24 h. The cells 
were fixed and permeabilized before fluorescent immunostaining and confocal 
microscopy (see Subheadings 3.3-3.6.)- Optical sections (350-nm thick) for a resorp- 
tion site were taken at the bone surface in the xy plane in (A-C) and through the center 
of the resorption site in the zx plane in (D) and are displayed in gray scale. The single 
fluorescent images in (A) show F-actin; in (B, D), lysosomal associated proteins; and 
in (C), the biotinylated dentine proteins. In (A) and (D) the resorption path is indicated 
by a broad arrow and the ruffled border is behind the broad arrow in (D). The sealing 
zone, ruffled border, and degraded bone matrix in resorbing osteoclasts are arrowed in 
(A), (B), and (C), respectively. The zx image in (D) was taken at the position identi- 
fied by "Section" in (A). In (D) the broken line, the open circle, and the open and 
closed stars show the position of the dentine surface, the sealing zone, the cell's apex, 
and cell base, respectively. A colored overlay of these images is shown in Fig. 5B, and 
C. (Reduced from original magnification, x630.) 


2/26/03, 10:47 AM 

I ■ I I I 


Nesbitt and Horton 



je cryosection 

* Cell top % 


": \J,\i 

i^ Cell base i% 


Side view ^ 

Fig. 4. High-power sectional views of osteoclasts in vivo in osteoclastoma tissue. A 
thick, 30-p.m, cryosection of osteoclastoma tissue was fixed and permeabilized and 
processed by fluorescent immunostaining for the metalloproteinase MMP-9, before 
confocal microscopy (see Subheading 3.8.)- Optical sections (350-nm thick) were 
taken through the tissue in the xy plane in (A) and in the zx plane in (B) and are 
displayed in gray scale. Cytoplasmic staining of MMP-9 is seen in the multinucleated 
osteoclasts in (A) and (B). The zx image in (B) was taken at the position in the tissue 
identified by "Section" in (A). Closed and open stars in (B) show the osteoclast's cell 
apex and cell base, respectively. A multicolored fluorescence intensity image of the zx 
section in (B) is shown in Fig. 5F. (Reduced from original magnification, x630.) 

Fig. 5. (opposite page) Multicolored imaging of osteoclasts by fluorescence 
immunostaining and laser scanning confocal microscopy. The cells were fixed and 
permeabilized before fluorescent immunostaining and confocal microscopy (see Sub- 
headings 3.3-3.6.)- In (A)-(D) human osteoclasts were cultured on biotinylated den- 
tine for 24 h. The pseudocolors of fluorescence represented are green (lysosomal 
associated membrane proteins), red (F-actin), and blue (dentine matrix proteins). Low- 
power and high-power views of resorbing osteoclasts are shown in (A, and B-D), 
respectively. In (A), many resorbing osteoclasts are identified by their characteristic 
F-actin staining of the sealing zones and form ring/arc-shaped structures. The resorp- 
tion sites are located by a loss of the dentine matrix proteins (black). In (B), optical 
sections of resorbing osteoclasts (350-nm thick) were taken at the bone surface in the 
xy plane (B, D) and through the center of the resorption site in the zx plane in (C). A 
pseudo-3D image in (D) shows the image in (B), rotated through 180° and tilted at an 
angle of 45°. The fluorescence intensity profile for the osteoclast in (B-D) is dis- 
played in (E), for each of the 16 optical sections taken through the resorption site. 
In (F), a fluorescence intensity image shows the localization pattern of the 
metalloproteinase MMP-9. Two osteoclasts are viewed from a zx section taken 


2/26/03, 10:47 AM 

I ■ I I I 

Confocal Microscopy of Osteoclasts 



A Above over vie 

Z-X section 

Seoling Cell apex Bone 
zone ;|; surface 


path Ruffled Cell base 

Side view 

a tt m *- tf « in 

E Fluorescence profile 

Tissue cryosection 

Side view io^m 

Fig. 5. (continued) through a thick cryosection of osteoclastoma tissue and show "hot 
spots" of MMP-9 colored white (colors white > red > yellow > green > blue > purple 
show the fluorescence intensities from high through to low). 



2/26/03, 10:47 AM 

I ■ I I I 


Nesbitt and Horton 


Fig. 6. Multicolored 3D imaging of bone-resorbing osteoclasts. Human osteoclasts 
were cultured on biotinylated dentine for 24 h. The cells were fixed before fluorescent 
immunostaining and confocal microscopy (see Subheadings 3.3-3.6.). The 3D 
isosurface images were reconstructed using Imaris software from 64 optical sections 
(each 350-nm thick) taken through a site of osteoclastic resorption in the xy plane. The 
pseudocolors of fluorescence represented are green (the integrin a v |3 3 , stained before 
cell permeabilization), red, and blue (F-actin and dentine matrix proteins, respectively, 
stained after cell permeabilization). The resorption path (broad arrow), the osteoclast's 
cell apex (closed star), and cell base (open star) and the edge of the bone surface 
(arrowed) are shown. In (A), an above view of a resorption site tilted at an angle of 30° 
shows the extracellular face of the plasma membrane of a resorbing osteoclast (green) 


2/26/03, 10:48 AM 

I ■ I I I 


Confocal Microscopy of Osteoclasts 275 

7. Take the images from steps 3 and 4 and display in a pseudo-3D format. Rotate 
and tilt the images and view from different perspectives (Fig. 5D). 

8. Use the gamma correction to visualize weak staining. Avoid boosting the gamma 
correction beyond 20% as nonspecific signals may be given (diffuse and 

9. Convert single fluorescence stains into a multicolored fluorescence intensity 
images and locate "hot spots" of staining with the glow-over facility (Glut mode) 
(Fig. 5F). 

10. Quantify the fluorescence stains. Map the area for analysis and measure the fluo- 
rescence intensity /jim 2 and display the fluorescence profile for the stack of im- 
ages (Fig. 5E). Analysis of the fluorescence intensities can be used to assess the 
amount of the resorption by osteoclasts (see Note 17). 

11. Finally, export the data from the 3D image collections (metSubheading 3.6.2., 
step 14) into a Silicon Graphics 2 Workstation (SGI, Mountainview, CA, USA) 
and generate 3D isosurface images using Imaris software (Bitplane, Zurich, Swit- 
zerland) (7,24). View the 3D images of the osteoclasts and resorption sites from 
above, laterally, and below the dentine surface (Fig. 6). 

3.7. Tracking of Exogenous Proteins in In Vitro Bone Resorption 

Many biological products, including peptides, proteins, and antibodies can 
be labeled effectively with a variety of fluorochromes and this allows them to 
be "traced" within live resorption cultures by fluorescence microscopy and 
LSCM (11,25-27). Fluoro-Link reactive dyes (Amersham/Pharmacia, UK) are 
convenient for labeling proteins and are nontoxic and stable in resorption 
assays. The fluorescence from Fluoro-Link reactive dyes is not quenched in 
acidic conditions and is ideal for imaging compounds that target the resorption 
site. These dyes allow compounds to be identified within their target cells and 
subcellular structures, and in resorbing osteoclasts include the resorption pit, 
ruffled border, sealing zone, basolateral cell body, and cell apex. 

3.7.1. Protein Labeling with FluorX 

1. Reconstitute the protein in 1 mL of 50 mM NaHC0 3 buffer, pH 8-8.5 (made 
fresh) and add to a FluoroX reaction vial. Mix thoroughly and avoid foaming of 
the solution. 

2. Cover the reaction vial with foil to shield from direct light and gently mix for 1 h at 4°C. 

3. After protein conjugation, remove the unbound FluorX by membrane filtration 
using Centricon devices. 

4. Aliquot and freeze the FluorX-labeled proteins at -70°C for long-term storage. 

Fig. 6. (continued) and stromal cells (red) within the resorption pit (black) and the 
matrix surface (blue). In (B), a side view shows the resorption pit (under the surface 
of the matrix) together with the concealed ruffled border (red). 

19/Nesbitt/259-282/F1 275 1 2/26/03, 10:48 AM 

I ■ I I I 


276 Nesbitt and Horton 

3.7.2. Cell Imaging with FluroX-Labeled Proteins 

1. Add the desired concentration of the FluorX-conjugated proteins to the in vitro 
resorption cultures as described in Subheading 3.3. 

2. Run a parallel culture and add an equivalent amount of unlabeled protein. 

3. Fix and permeabilize the cell cultures as noted in Subheading 3.3., step 8, and 
Subheading 3.5.1. 

4. Fluorescently immunostain the cells for F-actin and matrix proteins described in 
Subheading 3.5.2. and omit steps 3 and 5 for the primary and secondary anti- 
body staining. Identify the resorbing osteoclasts and resorption pits. 

5. Image the cells of interest by LSCM as noted in Subheading 3.6. View the 
FluoroX-labeled proteins using the FITC wavelength settings. 

6. Quantify the labeled proteins in the cell images using fluorescence intensity pro- 
files, Subheading 3.6.3., step 10. 

7. Analyze the effects on osteoclast resorption (see Note 18). 

8. Repeat steps 4-8 for the unlabeled protein and compare results with the FluoroX- 
conjugated protein. Establish if the results are specific to the exogenous protein 
and do not arise from any nonspecific effects of the label. If variations are seen, 
then repeat FluoroX labeling at lower concentrations and repeat the cell cultures 
and resorption analysis. 

3.8 Fluorescence Immunostaining and LSCM of Other Cell Types and 

In addition to the cells attached to bone and dentine discs, many other cell 
and tissue preparations can be studied by fluorescence immunostaining and 
LSCM techniques described in Subheadings 3.3-3.7. The cells can be seeded 
onto glass coverslips, or tissue cryostat sections can be attached onto glass 
slides. See, for example, the analysis of osteoclasts in vivo within a thick 
cryosection of osteoclastoma tissue probed for the metalloproteinase MMP-9 
(Figs. 4 and 5F). 

4. Notes 

1. Typical sources of dentine include elephant tusk and sperm whale and hippo- 
potamus teeth. Compared to bone, dentine lacks a Haversian system and is not 
remodeled. Its autofluorescence is low, which is very advantageous for cell 
imaging by fluorescence immunostaining and LSCM. Obviously this material is 
not routinely available, and legal sources need to be identified. In the UK, the 
Division of Customs and Excise (Heathrow Airport, London, UK) will make con- 
fiscated elephant tusk available for scientific research. Other sources to consider 
are natural history museums and zoos. 

2. Correct orientation of the dentine cutting is essential. The dentine discs represent 
transverse cuts so that any tubules within the dentine do not appear longitudi- 
nally. Longitudinal channels weaken the discs and make them prone to breakage 
during handling. Bone, in comparison to dentine, has an extensive Haversian 

19/Nesbitt/259-282/F1 276 1 2/26/03, 10:48 AM 

I ■ I I I 


Confocal Microscopy of Osteoclasts 277 

system and less mineral content and requires thicker cuts (> 200 u.m), and this 
elevates the autofluorescence. An increase in the background signals, in addition 
to osteoclasts being hidden from view within the channels, can severely restrict 
image analysis. Furthermore, discard unevenly cut dentine as "wedged-shaped" 
discs; these make high-quality image collection by LSCM difficult. 

3. Alternative, quicker methods for cleaning and sterilizing dentine and bone have 
been described using sonication and UV light (28). These procedures will degrade 
surface proteins on the dentine discs and may effect their subsequent biotin label- 
ing and, moreover, may alter cell attachment and resorption in the osteoclast 

4. Use biotinylation products that are stable, non-toxic, and utilize spacer arms to 
reduce steric hindrance. The biotinylation reagent biotinamidocaproate N- 
hydroxysuccinamide ester enables protein labeling via attachment to lysine or 
cysteine residues of amine groups with very low steric hindrance (29,30) and 
these labels do not effect cell attachment or resorption by osteoclasts (6). Fur- 
thermore, the biotin label is resistant to extremes of pH (3.0-11.0). It is not 
degraded by the acidic conditions (pH 4) within the resorption site at the ruffled 
border (31) and, subsequently, binds to streptavidin-fluorochromes used to probe 
the resorption cultures in immunostaining and LSCM analysis. 

5. A pH 8-8.5 provides optimal protein biotinylation and minimizes protein deg- 
radation of the dentine discs. The biotinylation reagent is susceptible to hydrolysis 
and is stored at 4°C; therefore leave the reagent at RT for 30 min before opening to 
avoid contamination with condensation. Discard the reagent 3 mo after opening. 

6. The discs must be completely dry before storage and kept sterile to preserve the 
biotin label. The label is stable for several years when stored under these conditions. 

7. Biotinylation methods are also suitable to surface label bone slices and enable the 
resorption sites to be imaged by fluorescence immunostaining and LSCM (32). 

8. The tissue digest can become gelatinous on cooling below RT and this can restrict 
the extraction of the osteoclastoma cells. If this occurs use more MEM-GPS to 
rinse the cells through the sieve, ensuring MEM-GPS is used at RT. 

9. Osteoclasts from a variety of sources can be studied by the bone and dentine 
resorption slice assay and imaged by fluorescence immunostaining and LSCM 
(5 ,6 ,20 .25-27 ,32) . The osteoclasts can be generated in in vitro cultures or iso- 
lated from the long bones of fetal and neonate limbs from human, rat, mouse, rab- 
bit, and chick species (see Part II, Culture of Osteoclasts). 

10. Keep the cells at 37°C when the samples are removed from tissue culture. 
Prewarm and pregas the washes and media and use gel "cold-packs" at 37°C to 
hold the tissue culture dishes. A constant temperature and pH will maintain 
osteoclast viability, attachment, and cell polarization of subcellular structures 
prior to cell fixation. 

11. The osteoclasts can be cultured for 2 wk to achieve extensive resorption of the 
dentine matrix. The resorption media should be replaced every 2 d to maintain 
cell viability. The biotin label on the dentine discs is stable in long-term tissue 
culture for further biochemical analysis. 

19/Nesbitt/259-282/F1 277 I 2/26/03, 10:48 AM 

I ■ I I I 


278 Nesbitt and Horton 

12. Paraformaldehyde and methanol are suitable cell fixatives for resorption cultures 
and preserve the actin cytoskeleton and tubulin networks, respectively (5,21). 
The paraformaldehyde fixative contains sucrose to stop osmotic fluxes that may 
cause nonspecific endocytosis which would alter cell morphology during the fixa- 
tion process. The MEM/fixation buffer mix stabilizes the localization of the cel- 
lular proteins that are calcium-dependent. Methanol fixatives also solubilize the 
cell lipids and puncture holes in the cell membranes and can cause some proteins 
to be lost from the cells. 

13. The fixed samples can be conveniently stored in freezing buffer at -70°C and are 
stable for several years. Freezing the samples does not affect the cell or matrix 
morphology, or the subsequent analysis of the cells by fluorescence 
immmunostaining and LSCM. 

14. The Triton X-100 detergent is used with an actin cytoskeleton stabilizing buffer 
at 4°C to preserve the cell's structure (5). Saponin can be used as an alternative to 
Triton X-100; it creates pores within the cell membrane by altering the lipid struc- 
ture. Generally, antibody access to the ruffled border and resorption site with 
saponin is less than that given with Triton X-100. Furthermore, 0.1% saponin is 
required in the subsequent antibody staining solutions and washes to maintain 
cell permeability. 

15. Fluorescence antifade solutions are essential to minimize the photobleaching of 
the fluorochromes conjugates used to immunostain the cells. The exposure times 
to fluorescence are extended by > 10-fold in the presence of antifade reagents. 
Thus, the samples can be examined for several minutes before loss of the fluores- 
cence signals. Use PBS-antifade mountants; the glycerol-based antifade 
mountants are not suitable for imaging osteoclasts on bone and dentine substrates. 

16. Crosstalk can occur with FITC and TRITC, and TRITC and CY5 fluorochromes 
due, in part, to an overlapping of their emission spectra. Generally, this is notice- 
able when the fluorescence staining from one fluorochrome is significantly 
greater than the other. For example, crosstalk may occur form excessive staining 
of the sealing zone in resorbing osteoclasts with phalloidin-rhodamine. To avoid 
crosstalk, a laser beam splitter enables specific light excitation of each fluoro- 
chrome to be individually excited. This system functions during the series collec- 
tion where optical slices are sequentially and separately gathered for each 
fluorescence stain. 

17. Changes in the percentages of fluorescence intensities for the stained cells and 
subcellular structures can be measured and changes in their distribution noted, 
and related to cell function. The loss of the biotin-labeled matrix from the dentine 
surface is a measure of osteoclast resorption and can be quantified by image 
analysis (6). In addition, the degraded matrix that is transported through the 
osteoclast during resorption can be stained with specific antibodies and quanti- 
fied in the osteoclasts after LSCM and analysis of the fluorescence intensities 
(11). Furthermore, the numbers of pits and osteoclasts (resorbing and non- 
resorbing types) may be counted by fluorescence microscopy (33). After the 
LSCM analysis, histochemical staining for tartrate-resistant acid phosphatase (see 

19/Nesbitt/259-282/F1 278 1 2/26/03, 10:48 AM 

I ■ I I I 


Confocal Microscopy of Osteoclasts 279 

the chapter by Flanagan and Massey, this volume, for details) or staining for 
toludine blue may also be used to identify and count the osteoclasts and pits. 
18. Methods to investigate bone-resorbing osteoclasts by live cell imaging techniques 
are under investigation. LSCM systems are required that provide high-resolution 
images while maintaining an environment in which the osteoclast retains its cell 
polarization and bone resorption activity. Generally, the rapid photobleaching of 
fluorochromes and the generation of hydrogen peroxide, caused by the scanning 
laser used in conventional confocal microscopy, reduces cell viability. Thus, these 
imaging systems are impractical for prolonged study of live osteoclasts in bone 
resorption cultures. Two-photon laser scanning microscopy (TPLSM) may cir- 
cumvent these problems and operates with minimal photobleaching and 
phototoxicity (34). TPLSM, together with the application of new fluorochrome 
labels for the bone matrix and the resorbing osteoclasts, will enable the dynamic 
process of bone resorption to be studied. 


1. Teitelbaum, S. L. (2000) Bone-resorption by osteoclasts. Science 289, 1504-1508. 

2. McKee, M. D. and Nanci, A. (1996) Microscopy of bone. Microsc. Res. Tech. 33, 

3. Lehenkari, P., Charras, G., Nesbitt, S., and Horton, M. (2000) New technologies 
in scanning probe microscopy for studying molecular interactions in cells. Expert 

4. Reynaud, K., Nogueira, R., Kurzawa, R., and Smitz, J. (2001) Confocal micros- 
copy: principles and applications to the field of reproductive biology. Folia 
Histochem. Cytobiol. 39, 75-85. 

5. Lakkakorpi, P. T., Helfrich, M. H., Horton, M. A., and Vaananen, H. K. (1993) 
Spatial organization of microfilaments and vitronectin receptor, av(33, in osteo- 
clasts. /. Cell Sci. 104, 663-670. 

6. Nesbitt, S. A. and Horton, M. A. (1997). Trafficking of matrix collagens through 
bone-resorbing osteoclasts. Science 276, 266-269. 

7. Nesbitt, S., Charras, G., Lehenkari, P., and Horton, M. (2000) Three-dimensional 
imaging of bone-resorbing osteoclasts: spatial analysis of matrix collagens, cathe- 
psinK, MMP-9 and TRAP by confocal microscopy. J. Bone Miner. Res. 15, 1219. 

8. Baron, R., Neff, L., Brown, W., Courtoy, P., Louvard, D., and Farquhar, M. (1988) 
Polarised secretion of lysosomal enzymes: co-distribution of cation-independent 
mannose-6-phosphate receptors and lysosomal enzymes along the osteoclast 
exocyotic pathway. /. Cell Biol. 106, 1863-1872. 

9. Horton, M. A, Nesbitt S. A., Bennett, J. H., and Stenbeck, G. (2001) Integrins and 
other cell surface attachment molecules of bone cells, in Principles of Bone Biol- 
ogy, 2nd edit. (Bilezikian, J. P., Raisz, L. G., and Rodan G. A., eds.). Academic 
Press, San Diego. 

10. Wucherpfennig, A., Li, Y., Stetler-Steveson, W., Rosenberg, A., and Stashenko 
P. (1994) Expression of 92 kD type IV collagenase/gelatinase B in human osteo- 
clasts. /. Bone Miner. Res. 9, 549-556. 

19/Nesbitt/259-282/F1 279 1 2/26/03, 10:48 AM 

I ■ I I I 


280 Nesbitt and Horton 

1 1 . Nesbitt, S. and Horton, M. (1999) Extracellular annexin II increases trafficking of 
matrix collagens through bone-resorbing osteoclasts to promote bone resorption. 
Calcif. Tissue Int. 64 (Suppl. 1), S35. 

12. Boyde, A., Ali, N. N., and Jones, S. J. (1984) Resorption of dentine by isolated 
osteoclasts in vitro. Br. Dent. J. 156, 216-220. 

13. Chambers, T. J., Revell, P. A., Fuller, K., and Athanasou, N. A. (1984) Resorp- 
tion of bone by isolated rabbit osteoclasts. /. Cell Sci. 66, 383-399. 

14. Arnett, T. D. and Dempster, D. W. (1986) Effect of pH on bone resorption by rat 
osteoclasts in vitro. Endocrinology 119, 119-124. 

15. Horton, M. A., Rimmer, E. F., Lewis, D., Pringle, J., Fuller, K., and Chambers, T. 
J. (1984) Cell surface characterisation of the human osteoclast: phenotypic rela- 
tionship to other bone marrow-derived cell types. /. Pathol. 144, 282. 

16. Atkins, G., Haynes, D., Graves, S., Evdokiou, S., Bouralexis, S., and Findlay, D. 
(2000) Expression of osteoclast differentiation signals by stromal elements of 
giant cell tumors. /. Bone Miner. Res. 15, 640-649. 

17. James, I., Lark, M., Zembryki, D., et al. (1999) Development and characterisation 
of a human in vitro resorption assay: Demonstration of utility using novel 
antiresorption agents. /. Bone and Miner. Res. 14, 1562-1569. 

18. Helfrich, M. H., Nesbitt, S. A., Dorey, E. L., and Horton, M. A. (1992). Rat osteo- 
clasts adhere to a wide range of RGD (Arg-Gly-Asp) peptide-containing proteins, 
including the bone sialoproteins and fibronectin, via a (33 integrin. /. Bone Min. 
Res. 7, 332-343. 

19. Horton, M. A. (2001) Integrin antagonists as inhibitors of bone resorption: impli- 
cations for treatment. Proc. Nutr. Soc. 60, 275-281. 

20. Helfrich, M. H., Nesbitt, S. A., Lakkakorpi, P. T., et al. (1996) |31 integrins and 
osteoclast function: involvement in collagen recognition and bone resorption. 
Bone 19, 317-328. 

21. Palokangas, H., Mulari, M., and Vannanen, K. (1997) Endocytotic pathway from 
the basal plasma membrane to the ruffled border membrane in bone-resorbing 
osteoclasts./. Cell Sci. 110, 1767-1780. 

22. Chavrier, P., Parton, R. G., Hauri, H. P., Simons, K., and Zerial, M. (1990) Local- 
ization of low molecular weight GTP binding proteins to exocytotic and endocy- 
totic compartments. Cell 62, 317-329. 

23. Horton, M. A, Townsend, P. A., and Nesbitt, S. A. (1996) Cell surface attachment 
molecules in bone, in, Principles of Bone Biology (Bilezikian, J. P., Raisz, L. G., 
and Rodan, G. A., eds.), pp. 217-230. Academic Press, San Diego. 

24. Guilak, F. (1994) Volume and surface area measurement of viable chondrocytes 
in situ using geometric modelling of serial confocal sections. /. Microsc. 173, 

25. Salo, J., Metsikko, K., Palokangas, H., Lehenkari, P., and Vaananen, H. K. (1996) 
Bone-resorbing osteoclasts reveal a dynamic divison of basal plasma membrane 
into two different domains. /. Cell Sci. 109, 301-307. 

26. Townsend, P. A., Vilanoova, I., Teti, A., and Horton, M. A. (1999) (31 integrin 
antisense oligodeoxynucleotides: utility in controlling osteoclast function. Eur. J. 
Cell Biol. 78, 448-496. 

19/Nesbitt/259-282/F1 280 1 2/26/03, 10:48 AM 

I ■ I I I 


Confocal Microscopy of Osteoclasts 28 1 

27. Stenbeck, G. and Horton, M. A. (2000) A new specialized cell-matrix interaction 
in actively resorbing osteoclasts. /. Cell Sci. 113, 1577-1587. 

28. Walsh, C. A., Carron, J. A., and Gallagher, J. A. (1996) Isolation of osteoclasts 
from human giant cell tumors and long-term marrow cultures, in Methods in 
Molecular Medicine: Human Cell Culture Protocols (Jones, G. E., ed.), Humana 
Press, Totowa, NJ. 

29. Nesbitt, S. and Horton, M. (1992) A nonradioactive biochemical characterisation 
of membrane proteins using enhanced chemiluminescence. Analyt. Biochem. 206, 

30. Hnatowitch, D. J., Virzi, F., and Rusckowski, M. (1987) Investigations of avidin 
and biotin for imaging applications. /. Nucl. Med. 28, 1294-1302. 

31. Silver, I. A., Murrills, R. J., and Etherington, D. J. (1988) Microelectrode studies 
on the acid microenvironment beneath adherent macrophages and osteoclasts. Exp. 
Cell. Res. 175, 266-276. 

32. Salo, J., Lehenkari, P., Mulari, M., Metsikko, K., and Vaananen, H. K. (1997) Removal 
of osteoclast bone resorption products by transcytosis. Science 276, 270-273. 

33. Lader, C, Scopes, J., Horton, M. and Flanagen, A. (2001) Generation of human 
osteoclasts in stromal cell-free and stromal cell-rich cultures: differences in 
osteoclast CDllc/CD18 integrin expression. Br. J. Haematol 112, 430-437. 

34. Squirrell, J., Wokosin, D., White, J., and Bavister, B. (1999) Long-term two pho- 
ton fluorescence imaging of mammalian embryos without compromising viabil- 
ity. Nat. Biotechnol. 17, 763-767. 

19/Nesbitt/259-282/F1 281 1 2/26/03, 10:48 AM 

I ■ I I I 




2/26/03, 10:48 AM 

I ■ I I I 



Bone Histomorphometry 

Shobna Vedi and Juliet Compston 

1. Introduction 

Histomorphometric examination of bone biopsies provides information on 
bone turnover, remodeling, and structure, which cannot be obtained from other 
investigative approaches such as bone densitometry and biochemical markers 
of bone turnover. Recently, there have been significant advances in 
histomorphometric techniques with the use of computer-assisted analysis and 
the development of sophisticated approaches to assessment of microstructure 
of bone. The application of these techniques has been particularly valuable in 
analyzing the cellular pathophysiology of different forms of osteoporosis and 
in determining the mechanisms by which drugs affect bone. In this chapter we 
review current methodology used in the preparation and histomorphometric 
assessment of histological sections of bone, with particular reference to its ap- 
plication in humans. 

2. Materials 

2.1. Biopsy 

1 . Trephine biopsy needle with internal diameter of at least 6 mm. 

2. Demeclotetracycline. 

3. Sedative (e.g., 5-10 mg of Midazolam) and local anesthetic (20 mL of 1% 

2.2. Tissue Processing and Embedding 

1. Heavy-duty microtome with tungsten carbide knife (Bright Instruments, 

2. Tissue processor (e.g., Shandon). 

3. Ethanol series 70%, 80%, 90%, 100%. 

4. LR White resin (store at 4°C). 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


20/Vedi/283-298/F1 283 1 2/26/03, 10:48 AM 

I ■ I I I 


284 Vedi and Compston 

5. DePeX. 

6. XAM mountant (BDH, Poole). 

2.3. Stains 

1. Toluidine blue: Dissolve 20 mg of toluidine blue in 10 mL of Mcllvane buffer 
(25 mL of 5 mM disodium hydrogen orthophosphate and 75 mL of 8 mM citric 
acid). Adjust to pH 4.2 with 1 M NaOH. 

2. von Kossa: 2.5% Silver nitrate (store at 4°C protected from light). 

3. van Gieson: Add 20 mL of 1% acid fuchsin in water to 80 mL of 1 M picric acid. 

2.4. Analysis 

We use a semi-automated system consisting of the following components: 

1. Olympus BHS-BH2 microscope. 

2. BH2-DA drawing attachment. 

3. Digicad digitizing tablet with LED point light source (Kontron Ltd.). 

4. HBO-100W/2 light source with EY-455 excitation filter for tetracycline labels. 

5. CCD camera to capture images for analysis of cancellous bone structure (1,2). 

Manual analysis can be done with any good transmitted light microscope 
using a graticule attachment for eyepiece (e.g., Zeiss Integrationsplatte 1) (see 
Note 1). 

3. Methods 

3.1 . Bone Biopsy 

The iliac crest is the standard site for a bone biopsy in humans. A sample 
can be obtained using either a vertical or transverse approach, the latter being 
favored by most investigators since it provides a biopsy with two cortices and 
intervening cancellous bone. The vertical biopsy, by contrast, contains only 
one cortex. Transiliac biopsies are obtained approx 2.5 cm below and behind 
the anterior superior iliac spine of the crest. There are several specially designed 
trephines available on the market; ideally, for bone histomorphometry, a tre- 
phine with an internal diameter of at least 6 mm should be used. 

Iliac crest biopsy is usually carried out as an outpatient procedure under 
mild sedation (e.g., 5-10 mg Midazolam) and local anesthesia (20 mL of 1% 
Lignocaine). It is associated with low morbidity; hematoma is the most com- 
mon complication, occuring in <1% of cases. 

Administration of two time-spaced doses of tetracycline prior to bone biopsy 
enables assessment of dynamic indices of bone turnover and should be per- 
formed whenever possible (3). Various regimens have been described, involv- 
ing a 10- to 14-d gap between the two administrations, bone biopsy being 
performed 3-5 d after the last dose. The regimen used in our laboratory is as 

20/Vedi/283-298/F1 284 I 2/26/03, 10:48 AM 

I ■ I I I 


Bone Histomorphometry 285 

Days 1 and 2: demeclotetracycline 300 mg twice daily. 

Days 3-12: no demeclotetracycline. 

Days 13 and 14: demeclotetracycline 300 mg twice daily. 

The biopsy is performed 3-5 d after the last dose. 

Side effects of demeclotetracycline are rare but include nausea and diarrhea. 
Skin rashes occasionally occur and, although usually mild, may be severe and 
exhibit photosensitivity. Different tetracycline compounds may differ with 
respect to the extent of their uptake at mineralizing fronts. 

3.2. Fixation and Processing of Biopsy (see Note 2) 

1. Fix the biopsy in 70% ethanol for a minimum of 48 h. 

2. Dehydrate by placing the biopsy in 80% ethanol for 3 d in a polypropylene vial 
on a rock and roller machine. 

3. Repeat using 90% ethanol for 3 d. 

4. Repeat using 100% ethanol for 3 d. 

5. Remove the biopsy from the ethanol, place centrally in the bottom of a rubber mold, 
and cover with approx 2 mL of LR White resin. Incubate at room temperature for 2 d. 

6. Replace with 2 mL of fresh LR White resin and incubate for a further 2 d. 

7. Replace with 2mL fresh LR White resin and incubate for 2 more days. 

8. Place in an oven at 60°C for 2-3 h until the resin hardens. 

3.3. Sectioning and Mounting 

Sections should be cut using a heavy-duty microtome and tungsten carbide 
knife. For bone histomorphometry, sections of 8 fxm and 15 ^im thickness are 
cut every 200 fxm at several levels throughout the biopsy, ensuring that each 
replicate section contains different trabeculae. The 15-[xm sections should be 
mounted immediately in XAM for tetracycline label analysis. These sections 
remain unstained to be viewed under 365-nm UV light. 

3.4. Staining 

A number of different staining procedures may be used to demonstrate the 
histological features of bone: 

a. Toluidine blue (1%, pH 4.2) is a metachromatic stain used for examination of 
bone cell morphology, identification of resorption cavities, mineralization fronts, 
and bone remodelling units. Polarized light microscopy can be used on toluidine 
blue stained sections for the measurement of wall width and for identification of 
resorption cavities (see below). 

b. von Kossa with van Gieson counterstain is used to differentiate osteoid (red) 
from mineralized bone (black). 

c. Solochrome cyanin R (4) is a red acid triphenylmethane dye used to stain basic 
and acidic proteins. It allows distinction between osteoid (orange) and mineral- 
ized bone (blue/grey) and the mineralization front (dark blue). 

20/Vedi/283-298/F1 285 1 2/26/03, 10:48 AM 

I ■ I I I 

286 Vedi and Compston 

d. Goldner's trichrome stain (5) can also be used to distinguis bhetween osteoid (red) 
and mineralized bone (green). Details can be found in Chapter 24 by van 't Hof et al. 

e. Villanueva stain (6) may be used for concomitant observation of osteoid and 
tetracycline labels. 

Details of the toluidine blue and von Kossa staining procedures follow. All 
staining is performed on free floating sections. 

1. Toluidine blue: Free float the 8-um thick sections in the stain for 2 min and wash 
twice with distilled water. Place the stained section on a glass slide, trim off 
excess resin, and blot dry. Air-dry the section, clear with inhibisol, and mount in 

2. Von Kossa: Place the 8-u.m thick sections in 2.5% silver nitrate solution in bright 
light for 3-4 h until they turn black. Wash twice in distilled water to remove all 
traces of silver nitrate. Fix in 2.5% sodium thiosulfate solution for 2 min and 
wash twice in distilled water. Immerse in van Gieson counter stain for 15 min 
and wash twice in distilled water. Flatten onto slides, air-dry and mount in DePeX. 

3.5. Theoretical Basis of Histomorphometry 

Bone histomorphometry has a number of limitations some of which are 
related to the restrictions imposed by a single-biopsy site and disease heteroge- 
-m- neity, whereas others reflect methodological problems and difficulties in iden- 

tification of some key features of remodelling. The conventional histological 
sections on which bone histomorphometry is performed are viewed as 2D 
images in which profiles of 3D structures are seen. To extrapolate 2D data to 
3D quantities, stereological formulae have to be applied. These are based on 
assumptions that sampling is unbiased and random and that the structure is 
isotropic (i.e. evenly dispersed and randomly oriented in space) (7). However, 
in the case of bone, these conditions are not totally fulfilled and the expression 
of bone indices as 3D quantities is thus subject to some error. In practice, 
histomorphometric data may be reported either as 2D or 3D quantities; although 
there will be a difference in the absolute values and accuracy depending on 
which approach is taken, it will not influence comparisons between patient 
groups or the diagnostic value of histomorphometry, provided that the approach 
adopted is consistent. 

3.6. Nomenclature 

All histomorphometric measurements are described according to the stan- 
dardized system proposed by ASBMR nomenclature committee (8). Table 1 
shows the referents used in bone histomorphometry and Tables 2 and 3 show 
the primary and derived indices of bone remodeling. All data are expressed in 
the format of source, the measurement, and the referent. The source is the struc- 
ture on which the measurement is made, for example, bone tissue or bone sur- 

20/Vedi/283-298/F1 286 1 2/26/03, 10:48 AM 

I ■ I I I 


Bone Histomorphometry 

Table 1 

Referents Used in Bone Histomorphometry 


Referent (3D/2D) 

Abbreviation (3D/2D) 

Bone surface/perimeter 
Bone volume/area 
Tissue volume/area 
Core volume/area 
Osteoid surface/perimeter 
Eroded surface/perimeter 
Mineralizing surface/perimeter 
Osteoblast surface/perimeter 
Osteoclast surface/perimeter 










Table 2 

Primary Histomorphometric Indices (Expressed as 2D Measurements) 




Bone area 
Osteoid area 
Osteoid perimeter 
Osteoid width 
Osteoblast perimeter 
Wall width 

Mineralizing perimeter 
Mineral apposition rate 
Eroded depth 
Eroded cavity area 
Eroded perimeter 
Osteoclast perimeter 
Erosion length 
Cavity number 
Osteoclast number 


O.Ar/T.Ar or O.Ar/B.Ar 











N.Cv/B.Pm or /T.Ar 


c ;< 







um 2 




No./mm or /mm 2 

cells/mm 2 

face. Measurements can either be primary or derived. The referent is area or 
perimeter in 2D and volume or surface in 3D terminology. Primary 2D mea- 
surements of perimeter, area, and number are expressed in terms of the amount 
of tissue examined and can be compared between subjects only when related to 
a common referent such as a clearly defined area or perimeter within a section. 
Absolute perimeter length and absolute area have no 3D equivalent but if 3D 



2/26/03, 10:48 AM 

I ■ I I I 


Vedi and Compston 

Table 3 

Derived Histomorphometric Indices of Bone Remodeling 





Adjusted apposition rate 



Mineralization lag time 



Osteoid maturation period 



Bone formation rate 


(j,m 2 /p,m/d or %/yr 

Activation frequency 



Remodeling period 



Formation period 



Active formation period 



Quiescent period 



Resorption period 



Reversal period 



Total period 



Trabecular width 



Trabecular number 



Trabecular separation 


[im or mm 

terminology is used the quantities are referred to as volume and surface, the 
absolute values being identical. The fourth type of primary measurement is num- 
ber, which cannot be extrapolated to 3D unless serial sections have been used. 

Width measurements can be converted to thickness by dividing the width 
value by 4/jt for isotropic structures. Values for width can be obtained by di- 
rect measurement or indirectly by calculation from area and perimeter. 

Bone histomorphometry is most commonly applied to cancellous bone, 
which has a higher bone turnover than cortical bone. Cancellous bone may be 
divided into corticoendosteal and mid-cancellous regions; definition of these 
regions is somewhat arbitrary but a consistent approach should be used where 
possible, for example, using fixed distances from the outer periosteum. Cortical 
bone has been largely neglected by histomorphometrists despite its predomi- 
nance in the skeleton and its importance as a determinant of bone strength and 
fracture risk. Nevertheless, application of histomorphometric techniques to cor- 
tical bone has been described in the rib, iliac crest, and femoral neck (9-11). 

3.7. Limitations of Histomorphometry 

Measurement variance associated with bone histomorphometry arises from 
a number of factors including intraobserver, interobserver, intermethod, and 
sample variation (12-14). Observer variation is mainly due to the subjective 
criteria used for identification of features such as osteoid seams, bone struc- 



2/26/03, 10:48 AM 

I ■ I I I 

Bone Histomorphometry 289 

tural units, and resorption cavities. In addition, the criteria for corticomedullary 
differentiation, staining methods used and the magnification at which measure- 
ments are made all contribute to variation. Many of these sources of variance can 
be minimized by the standardization of staining methods, corticomedullary delin- 
eation, and magnification. In addition, standardization of the criteria used to iden- 
tify histological features should be employed where possible; 3 ^im is used as the 
lower limit for the recognition of an osteoid seam width (15) and resorption cavi- 
ties are identified under polarized light by the presence of cut off collagen fibers at 
the edge of the cavity (16). Measurement of the tetracycline-labeled perimeter is 
also subject to substantial variation between observers, especially if old labels are 
present as a result of tetracycline administration in the past. Current 
histomorphometric techniques are limited by the lack of reliable markers for acti- 
vation and resorption. At present, indices related to these processes are used to 
calculate dynamic indices of bone turnover, based on the assumption that bone 
resorption and formation are coupled temporally and spatially and that bone 
remodeling is in a steady state; neither of these assumptions is likely to be tenable 
in the presence of treated or untreated osteoporosis (17). 

3.8. Primary Histomorphometric Indices 

-m- Primary measurements are summarized in Table 2 and discussed individu- 

ally below. These indices are measured directly and include area measurements 
(e.g., bone area), perimeter measurements (e.g., osteoid perimeter) and certain 
distance measurements (e.g., osteoid seam width and mean wall width) (see 
Note 3). In principle, all distance measurements can be measured either di- 
rectly at multiple locations or by indirect calculation from area and perimeter 
measurements. The direct method is preferable (15) and provides a frequency 
distribution, standard deviation, and a mean value, which are necessary for 
reconstructing the remodeling sequence. 

In our laboratory we perform the primary measurements of bone area, 
osteoid perimeter, and osteoid seam width on von Kossa stained secions on a 
minimum of 25 fields from three to six sections. 

1. Bone area (B.Ar/T.Ar): Bone area is the percent area occupied by calcified bone 
in relation to the total area. 

2. Osteoid area (O.Ar/T.Ar): Osteoid area is the percent area occupied by osteoid in 
relation to the total area. 

3. Osteoid perimeter (O.Pm/B.Pm): Osteoid perimeter is the percent of the bone 
surface occupied by osteoid in relation to the total bone perimeter (see Note 3). 

4. Osteoid width (O.Wi): Osteoid width is measured at four equidistant points or 
eight points on seams longer than 600 jj,m in length. A minimum of 20 seams per 
biopsy is measured on the same sections used for osteoid perimeter. All seams 
with a width of 3 jxm or more are included. 

20/Vedi/283-298/F1 289 1 2/26/03, 10:48 AM 

I ■ I I I 


290 Vedi and Compston 

5. Osteoblast perimeter (Ob/Pm/B.Pm): Osteoblast perimeter is the percent of the 
bone perimeter occupied by plump, cuboidal osteoblasts in relation to the total 
bone perimeter. 

6. Wall width (W.Wi): The mean width of completed bone remodeling units is mea- 
sured on toluidine blue stained sections viewed under polarized light at xl56 
magnification. A minimum of 25 bone packets is measured on each biopsy from 
between three and eight sections. 

7. Mineralizing perimeter (M.Pm/B.Pm): Mineralizing perimeter is assessed as the 
extent of bone perimeter that exhibits either single or double labels. Measure- 
ments are performed at xl56 magnification on a minimum of six 15-ixm thick 
sections using fluorescence microscopy (see Note 4). 

8. Mineral apposition rate (MAR). The mineral apposition rate is calculated as the 
mean distance between double labels divided by the time period between the 
administration of the two labels. Measurements are made at the midpoint of each 
label at approx four equidistant points along each double-labeled surface. A mini- 
mum of 20 labels from six sections is measured. The interlabel period is calcu- 
lated as the number of d between the midpoints of the two labeling dose regimens 
of tetracycline, (see Note 5). 

9. Maximum eroded depth (E.De.Max): This is defined as the maximum depth of 
eroded cavities measured interactively (see Note 6). 

10. Mean eroded depth (E.De): This is defined as the mean depth derived from mea- 
surements made at four equidistant points along the resorption cavity (see Note 6). 

11. Eroded cavity area (E.Ar): This is defined as the area of eroded cavities, mea- 
sured automatically (see Note 6). 

12. Eroded perimeter ( This is the percentage of bone perimeter occupied by 
eroded cavities (see Note 6). 

13. Osteoclast perimeter: This is the percentage of bone perimeter covered by osteo- 

14. Erosion length: Mean length of the base of the eroded cavities drawn by using the 
cursor, measured interactively. 

15. Cavity number: This is the number of cavities per millimeter of bone perimeter 
measured automatically (see Note 6). 

16. Cavity count/mm 2 : This is the number of cavities per square millimeter medul- 
lary area (trabecular bone + marrow) (see Note 6). 

17. Osteoclast number: This is defined as the number of osteoclasts in relation to the 
total area of the section (see Note 6). 

3.9. Calculation of Derived Indices 

Derived indices and their measurement units are summarized in Table 3 and 
discussed individually below. 

20/Vedi/283-298/F1 290 I 2/26/03, 10:48 AM 

I ■ I I I 


Bone Histomorphometry 291 

1. Adjusted apposition rate (Aj.AR): This represents the mineral apposition rate or 
the bone formation rate averaged over the osteoid surface. In a steady state and in 
the absence of a mineralization defect, the adjusted apposition rate is the best 
estimate available for osteoid or matrix apposition rate, as it can be assumed 
that they occur at the same rate (although not synchronously). It is calculated as 

Aj.AR = MAR x M.Pm / O.Pm 

It is clear from above equation that Aj.AR is usually less than MAR and cannot 
exceed it. 

2. Mineralization lag time (Mlt): The mineralization lag time is the interval between 
the deposition and mineralization of a given amount of osteoid, averaged over 
the life span of the osteoid seam. It is calculated as follows: 

Mlt = O.Wi/ Aj.AR 

3. Osteoid maturation time (Omt): Osteoid maturation time represents the interval 
between the onset of deposition and onset of mineralization of a given amount of 
osteoid at each bone forming site. It results from processes such as collagen 
crosslinking which are necessary before mineralization can occur. In humans Omt 
is shorter than Mlt and never exceeds it. It is calculated as follows: 

Omt = O.Wi / MAR 

4. Bone formation rate (BFR): Bone formation rate can either be expressed in terms 
of osteoid perimeter (Aj.AR) or in terms of bone perimeter (tissue-based bone 
formation rate, BFR / B.Pm ) or bone area. BFR expresses the rate of bone forma- 
tion per unit of bone surface and is calculated as: 

BFR / B.Pm = MAR x M.Pm / B.Pm 

5. Activation frequency (Ac.F): Activation frequency is a key determinant of bone 
mass in the adult skeleton and increased Ac.F is an important mechanism of bone 
loss in osteoporosis (18). Ac.F is the frequency with which a given site on the 
bone surface will undergo new remodeling. At present there are no in situ mark- 
ers of activation and so it has to be calculated indirectly, as the frequency with 
which a given site on the bone surface undergoes new remodeling, as follows: 

Ac.F = (1/Tt.P) or (BFR / B.Pm) / W.Wi) 

6. Remodeling periods: The remodeling period is the average duration of a single 
cycle of bone remodelling at any point on the bone surface and is the sum of 
resorption, reversal, and formation periods. The formation period is the mean 
time required to build a new bone structural unit and is divided into the active 
(FPa+) and the inactive (FPa-) formation periods. FPa- is a measure of the "off- 
time," which accounts for the discrepancy between the osteoid and mineralising 

20/Vedi/283-298/F1 291 I 2/26/03, 10:48 AM 

I ■ I I I 


292 Vedi and Compston 

perimeters that is not attributable to label escape. The formation period (FP) can 
be calculated as follows: 

FP =W.Wi/Aj.AR 
FPa+ = W.Wi / MAR 

FPa- = FP - FPa+ 

Quiescent period (QP), erosion period (EP), and reversal period (Rv.P) are calcu- 
lated as follows: 

QP =Q.Pm/B.PmxFP 
EP =E.Pm/B.PmxFP 

Rv.P = Rv.Pm/B.Pm x FP 

The total period is defined as the time between the initiation of two successive 
remodeling cycles and is the sum of Rm.P and QP: 

Tt.P = Rm.P + QP 

3. 10. Assessment of Bone Structure 

Both cortical and cancellous bone structure are important determinants of 
bone strength. Cortical geometry, porosity, and width, together with cancel- 
lous bone size, shape, connectivity, and anisotropy, determine the mechanical 
strength of bone. A number of different approaches to the study of bone struc- 
ture have been described, including direct measurement or calculation of tra- 
becular width, separation and density (19), strut analysis (1), assessment of the 
trabecular bone pattern factor (20) and star volume (21). 

1. Trabecular width, separation and number: These indices can either be measured 
directly or calculated, as follows: 

a. Trabecular width (Tb.Wi) = (2 x B.Ar)/B.Pm 

b. Trabecular number (Tb.N): Tb.N = (B.Ar/T.Ar)/Tb.Wi 

c. Trabecular separation (Tb.Sp): = (1/Tb.N) - Tb.Wi 

2. Strut analysis: This method is based on the topological classification of struts and 
the definition of nodes and termini (8). The bone section is viewed with a CCD 
camera mounted on a light box, allowing the whole section to appear within a 
single field of view (magnification x9). Images of the whole bone sections are 
captured on the video screen and segmented to yield a gray-level image of the 
section. These binary images are interactively edited to remove "bone dust" at 
the edges of the section and rejoin torn trabeculae, referring to sections under the 
microscope. The right and left corticomedullary junctions are defined automati- 
cally. The upper and lower boundaries are defined by the operator and the images 

20/Vedi/283-298/F1 292 I 2/26/03, 10:48 AM 

I ■ I I I 


Bone Histomorphometry 293 

are saved as "binim" files. These files are skeletonized to give a symmetric axis 
of the original bone profile. The computer automatically identifies trabecular 
nodes and termini and produces images with color-coded struts. The total length 
of all struts is calculated and the length of individual strut types is expressed as a 
percentage of the total length (see Note 7). The following indices are generated: 

Tm./Tm: Number of struts joining free ends (termini). 

Nd./Tm: Number of struts joining a node to a free end. 

Nd./Nd: Number of struts joining two nodes. 

Nd./ Lp: Number of struts forming part of a closed loop. 

Cx./Tm: Number of struts joining a free end to the cortex. 

Cx./Nd: Number of struts joining a node to the cortex. 

Cx./Cx: Number of struts joining cortex to the same cortex. 

Nd/Tm: Ratio of all nodes to free ends in a section. 

3. Trabecular bone pattern factor: This index is based on the relationship between 
convex and concave surfaces (30), convexity indicating poor connectivity and 
concavity being associated with a well-connected structure. Trabecular bone pat- 
tern factor (Tb.Pf) is assessed automatically by measuring the trabecular area and 
perimeter within the active region of the binary image before and after dilatation. 
Tb.Pf is calculated as follows: 

Tb.Pf = (PI -P2)/(A1 - A2) 

Where PI is the original perimeter, P2 is the dilated perimeter, Al is the original 
area, and A2 is the dilated area. Values of Tb.Pf may be influenced by the com- 
puter-based smoothing technique, the degree of dilatation used, and the magnifi- 
cation at which the measurement is performed. 

4. Star volume: The star volume is defined as the mean volume of all parts of an 
object that can be seen unobscured from a random point within the object (22) 
and can be applied to either the trabeculae (trabecular star volume) or the marrow 
(marrow space star volume). The application of this measurement was described 
by Vesterby using the vertical section technique and a cycloid test system, to 
assess trabecular structure in both vertebral (23) and iliac crest cancellous bone 
(21). This method involves generation of intercepts from random sampling points, 
which in turn are used to measure the true Euclidean distance between the grid 
point and the intercept (/„). This distance is measured using a simple computer 
algorithm and raised to the power of three (/„ 3 ). Marrow space star volume (mirr 3 ) 
is calculated as: 

^ Vspace = 4 X (jt/3) X (/„ 3 ) 

Values obtained for marrow space star volume are significantly influenced by 
biopsy size, especially in poorly connected cancellous bone. 

3. 1 1. Three-Dimensional Approaches 

A number of approaches are now available for the generation of 3D images 
of bone, including reconstruction of serial sections (24,25); scanning and ste- 

20/Vedi/283-298/F1 293 1 2/26/03, 10:48 AM 

I ■ I I I 


294 Vedi and Compston 

reo microscopy; volumetric, high resolution, and microcomputed tomography 
and magnetic resonance imaging (26,27). Although the in vivo application of 
these approaches is currently limited, they can provide information not avail- 
able from histological sections, such as connectivity, anisotropy, and trabecu- 
lar size and shape. 

4. Notes 

1 . Histomorphometric measurements were originally accomplished by superimpos- 
ing a set of lines and points, known as a grid or graticule, on the image. This was 
either inserted into the eyepiece of the microscope or drawn to a much larger 
scale on a flat surface onto which the image was projected. Several grids have 
been described. The Zeiss Integrationplatte 1 consists of 25 points arranged in a 
square array. The points are located at the crossings of short vertical lines with 
long horizontal lines. Random test line orientation can be achieved using alterna- 
tive hemispherical lines, or a cycloid test grid for vertical sections (28). Alterna- 
tively, random orientation of test lines may be achieved by random rotation of 
the graticule between each field of measurement (29). Several commercial mea- 
suring systems are now available; alternatively, in-house systems can be designed 
using relatively simple hardware. The field area to be measured is defined using 
a square etched on the eyepiece graticule. The square is mapped to correspond 
with an active area on the table and an active drawing area on a video monitor. 
The system is calibrated for each magnification used. Computerized semi-auto- 
mated methods for bone histomorphometry have now largely superseded the use 
of graticules, because they are less labor intensive and tedious for the operator 
and can perform complex measurements (e.g., strut analysis) that cannot easily 
be achieved by non computerized techniques. The method which we use com- 
bines a manually operated interactive drawing system and a computer for storing 
measurement data. Perimeter measurements are made by tracing with the cursor 
LED and distance measurements are made either by dotting with the cursor on 
either side of the structure at equally spaced points or by tracing the outline of the 
structure. For example, to measure a bone structural unit, the cement line and the 
outer mineralized bone surface are drawn separately using the cursor and the 
digitizing tablet. The distances between the two lines are calculated automati- 
cally at four equidistant points and stored as a mean of four measurements. 

2. The preparation of high-quality histological sections is an essential prerequisite 
for bone histomorphometry. The specimens are embedded in a hard resin to allow 
preparation of undecalcified sections to assess mineralization of bone. The resins 
most commonly used are methylmethacrylate and LR White resin. In our labora- 
tory we use LR White resin because of its lower toxicity and the shorter embed- 
ding time relative to methylmethacrylate. 

3. Assessment of osteoid: Measurements of osteoid perimeter and osteoid seam 
width are strongly influenced by the magnification and the stain used. At low 
magnification it is difficult to differentiate between the thin endosteal membrane 
covering the quiescent bone surface and the osteoid seam. Therefore, all seams 

20/Vedi/283-298/F1 294 I 2/26/03, 10:48 AM 

I ■ I I I 


Bone Histomorphometry 295 

less than one lamella (approx 3 u,m) are excluded. 

4. Interpretation of single tetracycline labels: The presence of single labels reflects 
the labeling escape phenomenon, caused by initiation of mineralization before 
the first label or its termination between administration of the two tetracycline 
doses (30). If only double-labeled perimeters are considered, therefore, the 
actively mineralizing perimeter will be underestimated and thus the double plus 
half the single labels are included in this measurement. 

Md.Pm/B.Pm (%) = dL.Pm + (0.5 x sL.Pm)/B.Pm 

dL.Pm = double-labeled perimeter 

sL.Pm = single-labeled perimeter 
Single labels can also reflect the switch from an active to resting state in a minor- 
ity of osteoid seams, that is, the "on/off" phenomenon. In cases where only single 
labels can be detected, it is suggested that the mineralising perimeter is expressed 
as half the single-labeled perimeter. 

5. MAR is used to derive several indices of bone formation, and therefore an accu- 
rate measurement is of importance. In biopsies in which there is no uptake of 
tetracycline the MAR and the derived indices are treated as missing data. In cases 
where only single labels can be detected, the finite lower limit of 0.3 jxm/d for 
MAR value is used (31). 

MAR (nm/day) = L.Wi/LP 

where L.Wi is the interlabel distance and LP is the labeling period. 

6. Assessment of bone resorption: Accurate measurement of resorption cavities is 
associated with a number of problems related to their identification and, in particu- 
lar, characterization of cavities in which resorption has been completed. Problems 
related to identification of resorption can be resolved to a certain extent by using 
polarized light microscopy to demonstrate cut off lamellae at an angle to the bone 
whereas the presence of osteoid in the cavity (16) indicates that resorption has been 
completed. The presence of osteoclasts within the cavity may also aid identifica- 
tion, particularly when histochemical techniques are used to demonstrate the pres- 
ence of tartrate resistant acid phosphatase (although this is not specific to 
osteoclasts) (32,33). However, there is currently no certain method of distinguish- 
ing between completed cavities without osteoid and those cavities in which the 
process of resorption has been terminated prematurely ("arrested resorption") (34). 
One method of direct measurement of resorption cavity size is to count the number 
of eroded lamellae beneath the trabecular surface in resorption cavities and to char- 
acterize the cavities according to the type of cells present (35). This method enables 
measurement of completed resorption cavity size but is technically challenging and 
has not been widely adopted. The computerized method developed by Garrahan et 
al. (36) enables a relatively rapid and more widely applicable quantitative assess- 
ment of several resorption cavity characteristics including mean and maximum ero- 
sion depth, length, area, and number of cavities. This method involves 
reconstruction of the eroded bone surface by a curve fitting technique (cubic spline); 

20/Vedi/283-298/F1 295 1 2/26/03, 10:48 AM 

I ■ I I I 


296 Vedi and Compston 

alternatively, reconstruction can be done manually (37). Inclusion of all cavities in 
the measurements results in underestimation of the size of completed cavities but 
does provide information about the distribution of cavity indices and enables 
measurement of the eroded surface. Using this method, cavities are identified on 
toluidine blue stained sections, viewed under polarized light at xl56 magnifica- 
tion and measured at x62.5, xl56, and x312 magnification according to the size 
of the cavity. Cavities with a depth of >3 \im are included. A minimum of 20 
resorption cavities is assessed per biopsy, from between two to six sections. 
7. Analysis of struts: Struts at the border intersecting the upper and lower bound- 
aries are not included in the analysis. Results can also be expressed as a percent- 
age of tissue area, as the trabecular area is measured automatically at the same 


1. Garrahan, N. J., Mellish, R. W. E., and Compston, J. E. (1986) A new method for 
the analysis of two-dimensional trabecular bone structure in human iliac crest 
biopsies. /. Microsc. 142, 341-349. 

2. Compston, J. E., Garrahan, N. J., Croucher, P. I., and Yamaguchi, K. (1993) Quan- 
titative analysis of trabecular bone structure. Bone 14, 187-192. 

3. Frost, H. M. (1969) Tetracycline-based histological analysis of bone remodelling. 
Calcif. Tissue Int. 3, 211-237. 

4. Matrajt, H. and Hioco, D. (1996) Solochrome cyanin R as an indicator dye of 
bone morphology. Stain Tech. 41, 97-99. 

5. Goldner, J. (1938) A modification of the Masson trichrome technique for routine 
laboratory purposes. Am. J. Pathol. 14, 237-243. 

6. Villanueva, A. R., Kujawa, M., Mathews, C. H. E., and Parfitt A. M. (1983) Iden- 
tification of the mineralization front: comparison of a modified toluidine blue 
stain with tetracycline fluorescence. Metab. Bone Dis. Rel. Res. 5, 41-45. 

7. Parfitt, A. M.(1983) The physiological and clinical significance of bone 
histomorphometric data, in Bone Histomorphometry: Techniques and Interpreta- 
tions (Recker, R., ed.), CRC Press, Boca Raton, FL pp. 143-224. 

8. Parfitt, A. M., Drezner, M. K., Glorieux, F. H., et al. (1987) Bone 
histomorphometry: standardization of nomenclature, symbols and units. /. Bone 
Min. Res. 2, 595-610. 

9. Frost, H. M. (1963) Mean formation time of human osteons. Canad. J Biochem. 
Physiol. 41, 1307-1319. 

10. Agerbaek, M. O., Eriksen, E. F., Kragstrup, J., Mosekilde, L., and Melsen, F. 
(1991) A reconstruction of the remodelling cycle in normal human iliac cortical 
bone. Bone Miner. 12, 101-112. 

11. Bell, K. L.,Loveridge, N., Power, J., Garrahan, N. J., Meggitt, B. F., and Reeve, J. 
(1999) Regional differences in cortical porosity in the fractured femoral neck. 
Bone 24, 57-64. 

12. de Vernejoul, M. C, Belenguer-Prieto, R., Kuntz, D., et al. (1998) Bone histo- 

20/Vedi/283-298/F1 296 1 2/26/03, 10:48 AM 

I ■ I I I 


Bone Histomorphometry 297 

logical heterogeneity in postmenopausal osteoporosis: a sequential 
histomorphometric study. Bone 8, 339-342. 

13. Chavassieux, P. M., Arlot, M. E., and Meunier, P. J. (1985) Intermethod variation 
in bone histomorphometry: comparison between manual and computerised meth- 
ods applied to iliac bone biopsies. Bone 6, 211-219. 

14. Wright, C. D. P., Vedi, S., Garrahan, N. J., Stanton, M., Duffy, S., and Compston, 
J. E. (1992) Combined inter-observer and inter-method variation in bone 
histomorphometry. Bone 13, 205-208. 

15. Vedi, S. and Compston, J. E. (1984) Direct and indirect measurements of osteoid seam 
width in human iliac crest trabecular bone. Metab. Bone Dis. Rel. Res. 5, 269-274. 

16. Vedi, S., Tighe J. R., and Compston, J. E. (1984) Measurement of total resorption 
surface in iliac crest trabecular bone in man. Metab. Bone Dis. Rel. Res. 5, 275-280. 

17. Compston, J. E. and Croucher, P. I. (1991) Histomorphometric assessment of tra- 
becular bone remodelling in osteoporosis. Bone Miner. 14, 91-102. 

18. Frost, H. M. (1985) The pathomechanics of osteoporosis. Clin. Orthop. Rel. Res. 
200, 198-225. 

19. Parfitt, A. M., Mathews, C. H. E., Villanueva, A. R., Kleerekoper, M., Frame, B., 
and Rao, D. S. (1983) Relationship between surface volume and thickness of iliac 
trabecular bone in aging and in osteoporosis. Implications for the microanatomic 
and cellular mechanism of bone loss. /. Clin. Invest. 72, 1396-1409. 

20. Hahn, M., Vogel, M., Pompesius-Kempa, M., and Delling, G. (1992) Trabecular 
bone pattern factor — a new parameter for simple quantification of bone micro- 
architecture. Bone 13, 327-330. 

21. Vesterby, A. (1990) Star volume of marrow space and trabeculae in iliac crest: 
sampling procedure and correlation to star volume of 1st lumbar vertebra. Bone 
11, 149-155. 

22. Serra, J. (1982) Image Analysis and Mathematical Morphology . Academic Press, 
London, UK. 

23. Vesterby, A., Gundersen, H. J. G., and Melsen, F. (1989) Star volume of marrow 
space and trabeculae of first lumbar vertebra: sampling efficiency and biological 
variation. Bone 10, 7-13. 

24. Odgaard, A. and Gundersen, H. J. G (1993) Quantification of connectivity in can- 
cellous bone, with special emphasis on 3-D reconstruction. Bone 14, 173-182. 

25. Gundersen, H. J. G., Boyce, R. W., Nyengaard, J. R., and Odgaard, A. (1993) The 
conneulor: unbiased estimation of connectivity using physical disectors under pro- 
jection. Bone 14, 217-222. 

26. Majumdar, S. and Genant, H. K. (1995) A review of the recent advances in magnetic 
resonance imaging in the assessment of osteoporosis. Osteoporos. Int. 5, 79-92. 

27. Genant, H. K., Engelke, K., Fuerst, T., et al. (1996) Non invasive assessment of 
bone mineral and structure; state of the art. /. Bone Miner. Res. 11, 707-730. 

28. Vesterby, A., Kragstrup, J., Gundersen, H. J. G., and Melsen, F. (1987) Unbiased stereo- 
logical estimation of surface density in bone using vertical sections. Bone 8, 13-17. 

29. Kragstrup, J., Melsen, F., and Mosekilde, L. (1982) Reduced wall thickness of 

20/Vedi/283-298/F1 297 I 2/26/03, 10:48 AM 

I ■ I I I 


298 Vedi and Compston 

completed remodelling sites in iliac trabecular bone following anticonvulsant 
therapy. Metctb. Bone Dis. Rel. Res. 4, 181-185. 

30. Frost, H.M. (1983) Bone histomorphometry: choice of marking agent and label- 
ling schedule, in Bone Histomorphometry ." Techniques and Interpretation (Recker, 
R., ed.), CRC Press, Boca Raton, FL, pp. 37-51. 

31. Foldes, J., Shih, M. S., and Parfitt, A. M. (1990) Frequency distribution of tetra- 
cycline based measurements: implication for the interpretation of bone formation 
indices in the absence of double-labelled surfaces. /. Bone Min. Res. 5, 1063-1067. 

32. Evans, R. A., Dunstan, C. R., and Baylink, D. J. (1979) Histochemical identifica- 
tion of osteoclasts in undecalcified sections of human bone. Miner. Electrolyte 
Metab. 2, 179-185. 

33. Burstone, M. S.(1959) Histochemical demonstration of acid phosphatase activity 
in osteoclasts. /. Histochem. Cytochem. 7, 39-41. 

34. Croucher, P. I., Gilks, W., and Compston, J. E. (1995) Evidence for interrupted bone 
resorption in human iliac cancellous bone. / Bone. Miner. Res. 10, 1537-1543. 

35. Eriksen, E. F., Gundersen, H. J. G., Melsen, F., and Mosekilde, L. (1984) Recon- 
struction of the resorptive site in iliac trabecular bone; a kinetic model for bone 
resorption in 20 normal individuals. Metab. Bone Dis. Rel. Res. 5, 235-242. 

36. Garrahan, N. J., Croucher, P. I., and Compston, J. E. (1990) A computerised tech- 
nique for the quantitative assessment of resorption cavities in trabecular bone. 
Bone 11, 241-246. 

37. Cohen-Solal, M. E., Shih, M-S., Lundy, M. W., and Parfitt, A. M. (1991) A new 
method for measuring cancellous bone erosion depth: application to the cellular 
mechanisms of bone loss in postmenopausal osteoporosis. /. Bone. Miner. Res. 6, 

20/Vedi/283-298/F1 298 1 2/26/03, 10:48 AM 

I ■ I I I 



Transmission Electron Microscopy of Bone 

Vincent Everts, Anneke Niehof, and Wouter Beertsen 

1 . Introduction 

Ultrastructural analysis of bone and other mineralized tissues such as calci- 
fied cartilage and dentin is essential for the understanding of the cell-cell/cell- 
matrix interaction, composition, and three-dimensional organization of these 
tissues. A wide variety of techniques have been introduced to process such 
tissues. This chapter describes a few methods to process mineralized tissues 
obtained from different sources. In addition, attention is paid to processing of 
cultured bone explants for electron microscopic analysis. 

2. Materials 

Prepare all solutions containing fixative in a fume hood and use gloves. All 
compounds are very toxic and most are volatile. 

2. 1. Fixative 

The fixative is 4% paraformaldehyde and 1% glutardialdehyde in 0.1 M 
sodium cacodylate buffer, pH 7.4) (see Notes 1 and 2). 

1 . Heat 200 mL of distilled water to 70°C. 

2. Dissolve 40 g of paraformaldehyde and add approx 2.5 g of sodium hydroxide 
pellets. Allow the solution to cool (the solution should be clear). 

3. Add 21.4 g of sodium cacodylate. 

4. Add 40 mL of 25% glutardialdehyde and fill up to 800 mL with distilled water. 

5. Adjust the pH to 7.4 with 1 N HC1 and fill up to 1000 mL. 

6. This solution should be stored at 4°C and new fixative should be prepared each 
week (see Note 1). 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


21/Everts/299-310/F1 299 I 2/26/03, 10:48 AM 

I ■ I I I 


300 Everts et al. 

2.2. Osmium and Ferrocyanide Postfixative 

The osmium and ferrocyanide postfixative consists of 1% osmium tetroxide 
and 1.5% potassium ferrocyanide [K 4 Fe(CN) 6 ).3H 2 0] in 0.1 M sodium cacody- 
late buffer, pH 7.4. Stock solutions should be stored at 4°C (see Note 3). 

1. 2% Os0 4 stock solution: Add 1 g of Os0 4 crystals (EMS, crystalline, highest 
purity 99.95%) to 50 mL of double-distilled water in a stoppered dark glass vial. 
Gently shake the solution until the crystals are dissolved. Store the solution in the 
tightly closed vial at 4°C. To avoid blackening of the solution the vial has to be 
thoroughly cleaned with acetone to remove lipids (osmium is an excellent fixa- 
tive for lipids!), washed in double-distilled water, and dried. Caution: Use gloves 
and avoid any contact with the skin! 

2. 0.2 M Sodium cacodylate buffer: Dissolve 42.8 g of sodium cacodylate in 900 mL 
of distilled water. Adjust the pH to 7.4 and add distilled water to a volume of 
1000 mL. 

3. 3% Ferrocyanide stock solution: Dissolve 3 g of potassium ferrocyanide in 0.2 M 
sodium cacodylate buffer. 

4. Prior to use, mix one volume of 2% Os0 4 solution with one volume of 3% ferro- 
cyanide solution. 

2.3. Osmium and Cacodylate Postfixative 

The osmium and cacodylate postfixative consists of 1% osmium tetroxide 
in 0.075 M sodium cacodylate buffer (see Note 3). 

1. 4% Osmium tetroxide stock solution: Dissolve 1 g Os0 4 of crystals in 25 mL of 
distilled water according to the method described in Subheading 2.2., step 1. 

2. 0.1 M Sodium cacodylate buffer: Dissolve 21.4 g of sodium cacodylate in 900 mL 
of distilled water, adjust the pH to 7.4, and add distilled water to a volume of 
1000 mL. 

3. Prior to fixation, mix one volume of the 4% Os0 4 solution with three volumes of 
0.1 M sodium cacodylate buffer. 

2.4. Decalcification Solution 

The decalcification solution consists of 1.9% glutaraldehyde and 0.15 M 
EDTA (Titriplex III, ethylenedinitrilotetraacetic acid disodium salt dihydrate) 
in 0.06 M sodium cacodylate buffer. 

1. Dissolve 38.53 g of sodium cacodylate and 167.52 g of Titriplex III in 2 L of 
distilled water. 

2. Stir the solution, and as soon as all Titriplex is dissolved (the solution should be 
clear) add 232 mL of 25% glutaraldehyde. 

3. Adjust the pH to 7.4, first by adding approx 10 g of sodium hydroxide pellets 
followed by adding 2 N sodium hydroxide. Add distilled water to a volume of 3 L 
(see Note 1). This solution is stable for several mo at 4°C. 

21/Everts/299-310/F1 300 1 2/26/03, 10:48 AM 

I ■ I I I 


Transmission Electron Microscopy of Bone 30 1 

2.5. Goldner's Masson Trichrome 

1. Dissolve 1.25 g of hematoxylin in 100 mL of 25% ethanol. 

2. Dissolve 0.15 g of Light Green SF Yellowish and 0.2 mL of glacial acetic acid in 
100 mL of distilled water. 

3. Ponceau de Xylidine stock: Dissolve 1 g of Ponceau de Xylidine and 1 mL of 
glacial acetic acid in 100 mL of distilled water. 

4. Acid fuchsin stock: Dissolve 1 g of acid fuchsin and 1 mL of glacial acetic acid in 
100 mL of distilled water. 

5. Ponceau-acid fuchsin stock: Two parts of Ponceau de Xylidine stock (item 3) 
with one part of acid fuchsin stock (item 4). 

6. Orange G stock: Dissolve 1 g of Orange G in 100 mL of distilled water. 

7. Ponceau-acid fuchsin staining solution: One part of Ponceau-acid fuchsin stock 
(item 5), one part of Orange G stock (item 6), and eight parts of distilled water. 

8. Dissolve 0.5 g of phosphomolybdic acid hydrate in 100 mL of distilled water. 

9. Dissolve 2.5 g of ferric chloride and 1 mL of concentrated HC1 in 99 mL of 
distilled water. 

10. Rinsing solution: 5.2 mL of 96% acetic acid in 1000 mL of distilled water. 

11. All staining solutions are stable for months and are stored at room temperature. 

2.6. Methylene Blue 

1 . Dissolve 2 g of methylene blue (Merck no. 1 . 1 5943) in 1 00 mL of distilled water 
(solution a). 

2. Dissolve 0.5 g of Azure II (Merck no. 1.0921 1) in 50 mL of distilled water (solution b). 

3. Dissolve 2 g of Borax (disodium tetraborat-10-hydrat) in 100 mL of distilled 
water (solution c). 

4. Mix solutions a:b:c = 2: 1 : 1 , and store at 4°C. The staining solution is stable for months. 

5. Filter just before use. 

2. 7. von Kossa 

1. Dissolve 0.5 g of silver lactate in 100 mL of double distilled water. 

2. Dissolve 0.5 g of hydrochinon in 100 mL of double distilled water. 

3. Dissolve 5 g of sodium thiosulfate pentahydrate in 100 mL of distilled water. 

4. All solutions are made fresh just before use. 

2.8. Uranyl Acetate 

1 . Dissolve 0.35 g of uranyl acetate in 10 mL of double-distilled water. Store at 4°C. 

2.9. Lead Nitrate 

1. Boil and cool 50 mL of double-distilled water. 

2. Dissolve 1.33 g of lead nitrate and 1 .76 g of trisodium citrate dihydrate in 30 mL 
of water. 

3. Shake vigorously for 1 min and then a few times during the following 30 min. 

4. Add 8 mL of 1 N NaOH and boiled water to a volume of 50 mL. Store at 4°C. 

21/Everts/299-310/F1 301 1 2/26/03, 10:48 AM 

I ■ I I I 


302 Everts et al. 

2. 10. Epoxy Resin 

Use gloves and a fume hood when preparing the stock of epoxy resin. 

1. Mix under continuous stirring the resin components (Ladd Res. Industries, 
Williston, VT), adding the next component when the previous one is completely 
dissolved. The components should be added in the following order: 100 g of LX- 
112 epoxy resin, 72.4 g of dodecenylsuccinic anhydride (DDSA), 40.4 g of 
nadicmethyl anhydride (NMA), and 3.9 g of 2,4,6-tri(dimethylamine-methyl) 
phenol (DMP-30). 

2. Stir very well for another 30 min and collect the mixed resin in small, 10-mL 
plastic vials with cap. 

3. Store these vials at -80°C. The frozen vials can be kept at this temperature for a 
very long time (at least for 1 yr). 

4. Prior to embedding, warm an appropriate number of vials at ambient tempera- 
ture. Open the dried vial only when the resin is at room temperature. 

3. Methods 

3. 1. Perfusion Fixation of Animal Bones 

1 . Use a perfusion fixation system consisting of a perfusion pump or a bottle with a 
rubber cap hanging upside down at a height of approx 50 cm above the work 
area. If a pump is used, the tube is inserted in a bottle with fixative. If a hanging 
bottle is used, a needle (0.8 x 40 mm) is fixed to a tube and inserted into the 
rubber cap of the bottle. In addition, inserting a second tube with a needle into 
the cap makes an air inlet. The other end of this tube is fixed to the side of the 
bottle, with its opening above the fluid level in the bottle. To the tube used for 
fixation a hypodermic needle (0.6 x 30 mm for small animals [e.g., young mice], 
0.8 x 40 mm for larger animals) is fixed for insertion into the heart. A valve 
should be placed somewhere along the length of the tube used for fixation. 

2. Anesthetize the animal and fix it on its back on a plateau. Open the belly and cut 
the thorax left and right from the sternum. Reflect the skin and open the thorax, 
expose the heart, and cut the pericardium. 

3. Fix the heart with two fingers (use well-fitting gloves) and insert the needle 
through the wall of the left ventricle. Open the tube valve (hanging bottle system) 
or switch on the perfusion pump (set to 2.5 mL/min). Wait a few seconds and cut 
the right atrium with a fine pair of scissors to let the perfusate escape. 

4. The quality of fixation is checked by testing the stiffness of soft tissues such as 
the lip, the bleaching of the liver, and rigidness of the paws. After 5-10 min the 
perfusion is stopped and the tissues are collected and stored in fixative. 

3.2. Immersion Fixation of Animal Bones 

1 . After euthanizing the animal and exposure of the bones of interest, dissect the bones 
and immerse them as quickly as possible in the fixative (4% paraformaldehyde and 
1 % glutaraldehyde in 0. 1 M sodium cacodylate buffer, pH 7.4). If bones are collected 
from larger animals, the bones should be cut into smaller pieces. Cutting is preferably 
done in fixative. (Bones of young mice can be fixed without further cutting.) 

21/Everts/299-310/F1 302 I 2/26/03, 10:48 AM 

I ■ I I I 


Transmission Electron Microscopy of Bone 303 

2. Fix at ambient temperature for at least 4 h. After this the tissue samples can be 
left in fixative overnight at 4°C. 

3. Transfer to postfixative for 1 h (see Note 3). 

4. Wash the sample in 0. 1 M sodium cacodylate buffer. 

5. Proceed with the embedding protocol (see Subheading 3.6. and following). 

3.3. Immersion Fixation of Human Bone Samples 

Immersion fixation and processing of bone samples obtained from humans 
is similar to the protocol described in Subheading 3.2. It is essential that the 
samples are immersed into the fixative as quickly as possible and that the size 
of the fragments is small. Try to keep a maximal thickness of approx 3-5 mm. 
Cutting of the bones into smaller fragments has to be performed in fixative. 

3.4. Immersion Fixation of Cultured Mineralized Tissues 

1 . Collect the bone explants after the preferred culture period and place in fixative 
(see Note 4). 

2. Leave at ambient temperature for at least 4 h (after this period the fixed bones can 
be kept at 4°C for another day). 

3. Process the bones further with or without decalcification (see Subheading 3.5.). 
Bones obtained from (very) young animals (e.g., mice <10 d) can be processed 
without the decalcification step. 

3.5. Decalcification of Mineralized Tissues 

1. Following fixation, immerse the bone samples in decalcification solution. 

2. Place the samples at 4°C for 2-3 wk, replacing the decalcification solution weekly. 

3. Check whether decalcification has been completed by X-ray photography. 

4. Once decalcification is complete, proceed to embedding. 

3.6. General Approach to Embedding and Analysis of Mineralized 
Tissues by TEM 

Special care has to be taken in the embedment of calcified tissues, as epoxy 
resins do not easily penetrate such tissues. Bones and/or teeth obtained from 
very young animals can be easily embedded in resin without giving problems 
with cutting and/or staining. Special attention has to be paid to embedding of 
bones from older and larger animals or man. 

3.7. Embedding of Small Tissue Samples 

Tissue samples obtained from young animals (up to 1 wk old) of approx 3- 
5 mm in thickness are put into glass (plastic dissolves in propylene oxide!) 
vials that can be closed and they are embedded as follows: 

1. Immerse the sample in 70% ethanol for 5 min. Repeat this three times. 

2. Replace with 80% ethanol for 5 min. Repeat this three times. 

3. Replace with 90% ethanol for 5 min. Repeat this three times. 

21/Everts/299-310/F1 303 1 2/26/03, 10:48 AM 

I ■ I I I 


304 Everts et al. 

4. Replace with 96% ethanol for 5 min. Repeat this three times. 

5. Immerse the tissue in 100% ethanol and close the bottle. Incubate for 10 min. 
Repeat the 10 min incubation in 100% ethanol three times in total. 

6. Replace the ethanol is replaced by propylene oxide. Incubate in a closed vial for 
10 min. Change the propylene oxide and repeat once. 

7. Dissolve the epoxy resin in propylene oxide at a 1:1 concentration. Immerse the 
tissue samples in the epoxy resin-propylene mixture, close the vials, and leave 
overnight with gentle shaking. 

8. Replace with pure epoxy resin and with the vials left open, place on a shaker and 
leave gently shaking for 5 h. 

9. Immerse the tissue samples in fresh epoxy resin in plastic molds, label, and leave 
overnight in an oven at 40°C. 

10. Transfer to an oven set at 60 C C to allow polymerization of the resin. 

3.8. Embedding of Larger Bone Samples 

It is possible to embed tissue samples of the size of a lower jaw of a full-grown 
mouse with good results. Larger samples have to be reduced in size or have to be 
cut in smaller pieces to obtain sufficient penetration of the epoxy resin. 

1. Follow the procedure as indicated under Subheading 3.7. but increase the dura- 
tion of each ethanol dehydration step to 10 min and each propylene oxide 
impregnation steps to 20 min. 

2. Immerse the samples in propylene oxide-epoxy resin at a ratio of 3:1 for 3 h. 

3. Immerse the samples in a 1:1 mixture of propylene oxide-epoxy resin for 3 h. 

4. Immerse the samples in a 1:3 mixture of propylene oxide-epoxy resin overnight. 

5. Immerse the samples in pure epoxy resin, with the vials open and shaking for 6 h. 

6. Embed the samples in fresh epoxy resin and leave overnight in an oven at 40°C. 

7. Proceed with polymerization as described in Subheading 3.7., step 10. 

3.9. Sectioning Mineralized Tissues 

Sectioning of the tissue is best performed using diamond knives, which are 
available both for semithin and ultrathin sectioning (see Notes 5 and 6). 

1. Use a glass knive to trim the block. 

2. Proceed with semithin and ultrathin sectioning using a diamond knife with the 
cutting angle set at 6°. 

3. Make sure the microtome to manually cut at a very low speed (1 mm/sec). 

4. To avoid wetting of the surface of the tissue block, keep the water level of the 
trough as low as possible. 

3. 10. Methylene Blue Staining of Semithin Sections 

The methylene blue staining solution stains most tissue components. It is an 
excellent stain for general purposes. Owing to its metachromatic properties, 
some components (e.g., cartilage and granules of mast cells) stain purple. 

21/Everts/299-310/F1 304 I 2/26/03, 10:48 AM 

I ■ I I I 

Transmission Electron Microscopy of Bone 305 

1. Cut semithin sections of approx 1-2 [im thickness. 

2. Collect the sections on a drop of water on a glass slide and dry the sections on a 
heating plate (60-70°C); leave the sections on the plate for an hour or longer. To 
avoid wrinkles of sections of larger tissue samples sections are best dried on a 
plate at 50°C. Dry these sections overnight before staining. 

3. When the sections are dry, add a drop of the filtered methylene blue staining 
solution to the section. 

4. Leave the staining solution for approx 15 sec (depending on the type of tissue 
and thickness of the section). 

5. Wash the section extensively with a jet of distilled water. 

6. Dry the washed section and cover it with a drop of epoxy resin. 

7. Cover the section with a cover glass and leave the section for several hours on a 
hot plate (or in a stove at 60°C) to polymerize the resin. 

3.11. Modified Goldner's Masson Stain for Sections of Epoxy 

The Goldner stain (1) stains nonmineralized bone (e.g., osteoid) red, miner- 
alized bone green, and calcified cartilage very light green. It also provides a 
very good staining for cells associated with bone and cartilage (Fig. 1). All 
staining procedures are carried out on a hot plate at 68-70°C. 

~W~ 1. Cut sections of approx 2.5 [im thickness. 

2. Collect the sections on a coated (e.g., Vectabond, Vector Laboratories, #SP-1800) 
slide and dry them on a hot plate. 

3. Wet the sections with distilled water and shake off excess of water. 

4. Stain for 3 min with ferric chloride on hot plate. 

5. Rinse quickly with warm tap water and dry on hot plate. 

6. Stain for 25 sec with hematoxylin on hot plate. 

7. Rinse quickly with warm tap water and dry on hot plate. 

8. Stain for 8 min with Ponceau-acid fuchsin staining solution on hot plate. 

9. Rinse quickly with 0.5% acetic acid. 

10. Stain for 3 min with 0.5% phosphomolybdic acid on hot plate. 

11. Rinse quickly with 0.5% acetic acid. 

12. Stain for 3-5 min with Light Green SF Yellowish on hot plate. 

13. Rinse quickly with 0.5% acetic acid. 

14. Air-dry and cover the sections. 

3. 12. von Kossa 

The von Kossa staining procedure (2) results in a black staining of mineral- 
ized tissue parts. If methylene blue is used as a counterstain, the other tissue 
components stain in different intensities of blue. 

1. Cut sections of 2 [im thickness. 

2. Collect the sections on coated slides (see Subheading 3.11., step 2) and dry them 
on a hot plate. 

21/Everts/299-310/F1 305 1 2/26/03, 10:48 AM 

I ■ I I I 


Everts et al. 


Fig. 1. Light micrograph of a section of mouse calvarial bone stained with the 
Goldner Masson's Trichrome staining. The mineralized bone (B) is green and the 
osteoid (OS) is red. Section was obtained from epoxy resin embedded undecalcified 
bone. Bone was fixed in 4% paraformaldehyde and 1 % glutaraldehyde in 0. 1 M sodium 
cacodylate buffer. OB, Osteoblast. (Reduced from original magnification, x3000.) 

3. Incubate the sections with 0.5% silver lactate for 20 min at ambient temp. 

4. Rinse the sections with double-distilled water. 

5. Incubate the sections with 0.5% hydrochinon for 2 min at ambient temperature. 

6. Rinse the sections in double-distilled water. 

7. Incubate the sections with 5% sodium thiosulfate pentahydrate for 2 min at ambi- 
ent temperature. 

8. Rinse the sections in double-distilled water. 

9. Sections can be counterstained with methylene blue (see Subheading 3.10.)- 
10. Dry the sections and cover them. 

3.13. Staining of Ultrathin Sections with Uranyl Acetate 

1. Centrifuge the uranyl solution (10 min, 3000 rpm). 

2. Put drops of uranyl acetate on a strip of parafilm. 

3. Float the grids on top of the drops (sections facing the solution). 

4. Stain the sections for 4-8 min in the dark. 

5. Rinse the sections extensively with double-distilled water. 

6. Air-dry the sections and stain them with lead nitrate. 

3. 14. Staining of Ultrathin Sections with Lead Nitrate (3) 

1. Centrifuge the lead solution (10 min, 3000g). 

2. Put drops of lead nitrate on a strip of parafilm. 



2/26/03, 10:48 AM 

I ■ I I I 


Transmission Electron Microscopy of Bone 307 

3. Place a few sodium hydroxide pellets around the drops. 

4. Float the grids on top of the drops of the lead solution. 

5. Cover the drops and pellets with a cover of a Petri dish. 

6. Stain the sections for 2-4 min. 

7. Rinse the sections extensively with double-distilled water. 

8. Air-dry the sections. 

4. Notes 

1. The fixatives mentioned in this chapter give reproducible and reliable results 
(Figs. 2 and 3). The recipes given are for large volumes. If your laboratory does 
not have a high throughput of samples, the amounts can be reduced appropriate 
to your needs. One of the advantages is that tissue samples can be stored for a 
very long time in the aldehyde mixture without noticeable loss of ultrastructural 
details. Essential is that the paraformaldehyde is not older than 1 wk before its 
use for the initial fixation of tissue samples. 

2. Apart from the fixative mentioned in this chapter numerous others mentioned in 
the literature could be used that may give very good results. Caution: Whatever 
fixative used, it is crucial to avoid the use of phosphate buffer! Fixation carried 
out in phosphate-buffered fixatives, in particular if the fixative is somewhat older, 
results in formation and precipitation of mineral crystallites (Fig. 4). These crys- 
tallites are formed at the conjunction of mineral-containing tissue (e.g., bone) 
and the surrounding soft tissue but also within the cells, in particular in vacuoles 
and mitochondria. Owing to a relatively high calcium concentration at these sites 
calcium-phosphate crystals are rapidly formed. 

3. Postfixation with osmium-ferrocyanide results in strongly contrasted plasma 
membranes. If this is not desired, postfixation should be carried out in osmium 
without ferrocyanide (see Subheadings 2.2. and 2.3.). 

4. Considerations mentioned under Note 2 are also relevant for the choice of medium 
used to culture bone explants. Different commercially available culture media con- 
tain different concentrations of phosphate; the higher the concentration level the 
higher the chance for precipitation of crystals as soon as calcium is liberated. Dur- 
ing our investigations on bone resorption, we have experienced this. We found that 
blocking the activity of certain proteolytic enzymes resulted in the occurrence of 
large areas of nondigested demineralized bone matrix adjacent to osteoclasts. This 
was clearly shown for bone explants cultured with M199 (4). By using other media 
(e.g., BGJb medium) precipitation of crystals may occur at these sites. 

5. A problem recognized for a long time is the loss of mineral due to handling and 
cutting of mineralized tissue. One has to be aware that during cutting of sections 
and their collection on water, mineral may easily dissolve. This is particularly the 
case if sections are collected on water with a low pH. Landis and co-workers (5) 
suggested using other techniques to overcome this problem. 

6. Mineralized tissues are characterised by a high electron density in mineral-con- 
taining parts of the sections. Yet, quite often it appears as if mineral is absent due 
to the absence of a strong electron density in such sections (Fig. 5). This problem 

21/Everts/299-310/F1 307 I 2/26/03, 10:48 AM 

I ■ I I I 


Everts et al. 


Fig. 2. Low-power electron micrograph of a mouse metatarsal fixed in 4% 
paraformaldehyde and 1% glutaraldehyde in 0.1 M sodium cacodylate buffer. OC, 
osteoclast; M, mineralized cartilage; asterisk indicates an area where the bone is absent 
due to ultrathin cutting of the mineralized matrix. (Reduced from original magnifica- 
tion, x5,000.) 

Fig. 3. Electron micrograph of osteoblasts (OB) separated from mineralized bone 
(B) by a layer of osteoid (OS). Calvarial bone of a 10-d-old mouse was fixed in 4% 
paraformaldehyde and 1% glutaraldehyde in 0.1 M sodium cacodylate buffer. 
Undecalcified bone. (Reduced from original magnification, x 16,000.) 



2/26/03, 10:48 AM 

I ■ I I I 

Transmission Electron Microscopy of Bone 



Fig. 4. Electron micrograph of a section of undecalcified bone obtained from a 
patient suffering from osteopetrosis. Note the presence of large crystallites (arrows) 
due to a phosphate-buffered fixative. Bone tissue was fixed in 4% paraformaldehyde 
and 1% glutaraldehyde in 0.1 M phosphate buffer. B, Bone; CYT, cytoplasm; N, 
nucleus. (Reduced from original magnification, x26,500.) 

Fig. 5. Electron micrograph of a section of undecalcified bone obtained from a 
control patient. Note the electron translucent area (asterisks) in the bone (B) part of 
the section. A high electron density (see area indicated by B in Figs. 3-4) character- 
izes a thicker section of the same area. Bone sample was fixed in 4% paraformalde- 
hyde and 1% glutaraldehyde in 0.1 M sodium cacodylate buffer. N, Nucleus; CYT, 
cytoplasm. (Reduced from original magnification, x 15,500.) 

is often due to the thickness of the section; the thinner the section the lower the 
electron density will be. If one wishes to compare the same area in a section with 
and without mineral, it is possible to decalcify ultrathin sections. To accomplish 



2/26/03, 10:48 AM 

I ■ I I I 


Everts et al. 

this, sections collected on grids are floated on a drop of 0. 1 M EDTA in buffer for 
10 min. The sections are washed, counterstained, and examined. In sections thus 
treated most mineral is dissolved but decalcification of mineral enclosed by epoxy 
resin (e.g., free in vacuoles) proves to be very difficult. 


1. Gruber, H. E. (1992) Adaptations of Goldner's Masson trichrome stain for the 
study of undecalcified plastic embedded bone. Biotech. Histochem. 67, 30-34. 

2. Rungby, J., Kassem, M., Eriksen, E. F., and Danscher, G. (1993) The von Kossa 
reaction for calcium deposits: silver lactate staining increases sensitivity and 
reduces background. Histochem. J. 25, 446-451. 

3. Reynolds, E. S. (1963) The use of lead nitrate at high pH as an electron opaque 
stain in electron microscopy. /. Cell Biol. 17, 208. 

4. Everts, V., Korper, W., Jansen, D. C, et al. (1999) Functional heterogeneity of 
osteoclasts: matrix metalloproteinases participate in osteoclastic resorption of cal- 
varial bone but not in resorption of long bone. FASEB J. 13, 1219-1230. 

5. Landis, W. J., Paine, M. C, and Glimcher, M. J. (1978) Use of acrolein vapors for 
anhydrous preparation of bone tissue for electron microscopy. /. Ultrastruct. Res. 
70, 171-180. 




2/26/03, 10:48 AM 

I ■ I I I 



Scanning Electron Microscopy of Bone 

Deborah Marshall, Miep H. Helfrich, and Richard M. Aspden 

1 . Introduction 

Scanning electron microscopy (SEM) is a technique whereby both struc- 
tural and analytical information can be obtained from bone. Ways to use SEM 
to gain information on bone remodeling and bone pathology have been dis- 
cussed in a number of comprehensive reviews (1-4). In contrast to transmis- 
sion EM, in which only very small pieces of tissue can be examined, the sample 
size for SEM is much less restrictive. Using SEM, a wide range of magnifica- 
tions (10- to 10,000-fold) can be employed to obtain good overviews as well as 
detailed images. 

SEM has been used extensively to analyze bone structure, to study bone 
resorption in vivo and in vitro, and to examine surface structures and cell- 
matrix interaction of bone cells with various substrates using standard second- 
ary electron imaging (5-8). Preparation techniques required are not different 
from those for nonmineralized tissues and a routine method is given in this 
chapter. SEM is very useful to obtain structural information about bone matrix. 
To examine the matrix, generally all cells and organic material are removed 
from the surface of the bone. Specimens can then be examined using secondary 
electron (SE) imaging, but use of backscattered electron (BSE) imaging has also 
been described and yields important additional qualitative information (9). 

BSE imaging has been used most extensively to obtain quantitative infor- 
mation (9-13). To obtain quantitative information it is critical to embed speci- 
mens in resin and highly polish their surface. This ensures that differences in 
surface roughness do not contribute to the BSE signal obtained. Specimens can 
be rendered anorganic before embedding (using the procedure given in this 
chapter), but other publications (11-13) describe BSE imaging on routine 
pathological samples embedded in polymethylmethacrylate (PMMA). 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


22/Aspden/311-320/F1 311 1 2/26/03, 10:48 AM 

I ■ I I I 


312 Marshall et al. 

Quantitative BSE imaging can yield detailed information on bone mineral- 
ization density distributions, which are independent of bone volume, and there- 
fore provide additional information to traditional bone mineral density (BMD) 
measurements using dual energy X-ray absorptiometry (DXA) and bone 
histomorphometry. Description of the detailed analytical methods is beyond 
the scope of this chapter, which concentrates on specimen preparation tech- 
niques only, and the reader is referred to recent publications describing in de- 
tail the application of these methodologies to hard tissues (10,11,13). 

Quantitative information from bone can also be obtained using electron 
probe microanalysis, which measures local elemental composition with a spa- 
tial resolution of several cubic micrometers (14). This allows, for example, the 
ratio of calcium to phosphorus to be measured in different sites (9). Specimen 
preparation is as described for qualitative BSE imaging. 

2. Materials 

2. 1. Equipment 

1. l-L Dewar flask. 

2. Insulated forceps. 

3. Sharpened chisel. 

4. 60°COven. 

5. 37°C Incubator. 

6. Sputter Coater, for example, Baltec SCD 030. 

7. Critial point dryer, for example, Baltec CPD 030. 

8. Scanning electron microscope (we used Jeol 35 CF and Philips XL 20). 

9. Aluminum stubs to hold specimen (Agar Aids). 

10. Silver Dag adhesive (Agar Aids). 

11. Double-sided carbon tape (Agar Aids). 

12. Desiccator. 

13. Bone saw, for example, Accutom 2 (Struers Ltd., Glasgow, UK), or Isomet 
(Buehler, Lake Buff, IL, USA). 

14. Lapping machine for polishing bone slices (Struers). 

15. Vacuum oven. 

2.2. Reagents 

1 . 0.1 M Phosphate buffer, pH 7.2. 

2. 0.089 M Phosphate buffer, containing 2.5 mM MgCl 2 , pH 7.2. 

3. Glutaraldehyde (25%, Agar Aids, or other). 

4. Osmium tetroxide (Agar Aids). 

5. Alcohol series (70% through to 100% dry ethanol). 

6. Acetone. 

7. 1 mg/mL of Proteinase K in 10% sodium dodecyl sulfate (SDS) solution in dis- 
tilled water. 

22/Aspden/311-320/F1 312 I 2/26/03, 10:48 AM 

I ■ I I I 


SEMofBone 313 

8. Diethyl ether. 

9. Distilled water. 

10. Liquid N 2 . 

11. LR White resin. 

3. Methods 

3.1. Cell Cultures 

This is a method for the preparation of bone cell cultures on coverslips or 
bone/dentine slices used in resorption assays. Careful preparation results in 
excellent cell and bone surface morphology. The morphology and depth of the 
resorption pits can give important clues about the ability of osteoclasts to resorb 
under different conditions and pharmacological treatments (see, for example, 
Fig. 1 and ref. 8). SEM was used extensively to quantify bone resorption in the 
pit assay before light microscopical systems were introduced (as described in 
Chapter 1 1 by van 't Hof, this volume). Quantification or examination of 
resorption pits is best performed on bone slices from which all cells have been 
removed, either manually by rubbing the slices between fingers, or by immers- 
ing slices in 0.2 M NH 4 OH or 5% sodium hypochlorite (for 1 min or longer as 
necessary). Cells are usually removed after the fixation step (step 2 below). 
Alternatively, slices are first examined with adherent cells after which cells are 
removed, slices recoated and reexamined to visualize the pits alone. 

1. Rinse cell cultures on dentine/bone slices, or coverslips in phosphate buffer, 
pH 7.2 (see Note 1). 

2. Fix in 2.5% glutaraldehyde, 2.5 mM magnesium chloride in 0.089 M phosphate 
buffer, pH 7.2 for 3 h. 

3. Rinse in 0. 1 M phosphate buffer. 

4. Postfix in 1 % osmium tetroxide for 1 h. 

5. Wash in distilled water three times. 

6. Dehydrate in a graded series of ethanol solutions (70% through to 100% dry 

7. Carefully remove dentine/bone slices, or coverslips from culture dishes using 
watchmaker's forceps. 

8. Critical point dry from liquid C0 2 . 

9. Mount slices onto aluminum stubs using the silver adhesive and check that they are 
mounted the correct way up using a binocular microscope to observe cell surface. 

10. Coat specimens with 20 nm of platinum (see Note 2). 

1 1 . Examine specimens in a scanning electron microscope at an accelerating voltage 
of 10 kV. 

12. Figure 1 shows and example of a bis-phosphonate treated osteoclast on a dentine 
slice, prepared using this method. Further examples can be found in the chapter 
by Coxon et al. and Chapter 6 by Collin-Osdoby et al., this volume. 

22/Aspden/311-320/F1 313 1 2/26/03, 10:48 AM 

I ■ I I I 


Marshall et al. 


Fig. 1. (A) Control osteoclast (arrow) demonstrating typical surface morphology 
with numerous small microvilli, in contrast to the smooth surface of the fibroblasts 
(asterisks). The osteoclast has just moved away from the resorption pit, which shows 
a characteristic fringe of demineralized collagen fibers. (B) Osteoclast treated with the 
bisphosphonate clodronate. Note the cell shows extensive surface blebs, indicative of 
apoptosis, and is not resorbing. Scale bar = 10 jxm. 

3.2 Preparation of Bone Cores/Bone Biopsies forSEM Using SE or 
BSE Imaging (see Note 3) 

SEM is extremely useful to examine the bone matrix itself. An important 
difference from the method described in Subheading 3.1. is all organic mate- 
rial is removed from the specimen. Because there is no need to preserve the 



2/26/03, 10:48 AM 

I ■ I I I 

SEM of Bone 



Fig. 2. Piece of subchondral bone from a femoral head of a patient with osteoarthri- 
tis. (A) All cells and organic surface material have been removed using the procedure 
described in Subheading 3.2. and the remaining mineralized bone was imaged using 
SE. (B) Same bone as in (A), examined with BSE. Because contrast in BSE imaging is 
largely influenced by surface topography, only the surface area of the bone should be 
considered and compared with the same surface as seen in the SE. With BSE heavily 
mineralized areas appear white, whereas poorly mineralized areas appear dark. Note 
how the subchondral plate is poorly mineralized. Also note how this area is not specifi- 
cally distinct in the SE image, scp, subchondral plate; b, bone. (Reproduced from: Li, B., 
Marshall, D., Roe, M., and Aspden, R. M. (1999) The electron microscope appearance 
of the subchondral bone plate in the human femoral head in osteoarthritis and osteoporo- 
sis. /. Anat. 195, 101-1 10; with permission from Blackwell Science Ltd.) 

integrity of the cellular component of bone, specimens do not have to be fixed 
and they may be stored frozen for considerable time (see Note 4). The method 
given here has been designed to be a nonaggressive means of removing all 
cellular and surface organic material from pieces of bone, to allow a detailed 
study of the bone structure, for example, the connectivity of trabecular bone, or 
the porosity of cortical bone. The resulting samples can be observed using con- 
ventional SE, or using qualitative BSE (Fig. 2). For quantitative analysis using 
BSE, the surface roughness of specimens should be minimal and embedding in 
plastic, followed by polishing the surface, is essential. A method for embed- 
ding bone in LR White is given here, but other methods, such as embedding in 
PMMA {see the chapter by van Leeuwen and Derkx, this volume), can also be 
used (10-13) and allow retrospective studies of routine pathological specimens. 

3.2.1. Removal of Organic Material from Bone Samples 

1. Obtain human samples of bone from orthopaedic surgery or postmortem exami- 
nation, or from animal bone. For storage see Note 4. 

2. There are a variety of methods to obtain samples of the required size (generally a 
few millimeters in any direction). For cancellous bone a dissecting knife or even 
a scalpel may be sufficient to cut through the tissue. Alternatively, a hacksaw 



2/26/03, 10:48 AM 

I ■ I I I 


316 Marshall et al. 

will produce rough samples. Precision cutting is done using a mineralogical saw 
fitted with either a diamond or an aluminum oxide cutoff wheel. We use an 
Accutom 2 (Struers) and a 125-mm aluminum oxide wheel (aluminum oxide in a 
medium-hard Bakelite bond) rotating at 600-800 rpm. Faster wheel speeds and 
slower specimen feed result in a better surface finish if this is required (see Note 
2 of Chapter 27 by Aspden, this volume). Perhaps the most common saw cited in 
the bone literature is the Isomet (Buehler). These saws are precision machines 
and therefore have a limited capacity. Cutting specimens roughly to size first is 
essential. An alternative, which produces surfaces that may be indicative of mate- 
rial properties, is to freeze-fracture. To do this on subchondral bone plate samples 
about 1-2 mm thick, samples are immersed in liquid nitrogen using insulated 
forceps and then fractured by bending and snapping (see Note 5 and Fig. 2). 

3. To remove organic material, place bone fragments in a screw-topped container 
containing 1 mg/mL of proteinase K and 10 mg/mL of SDS in distilled water. 
Place in an incubator at 37°C and leave overnight with gentle agitation. 

4. Change proteinase K-SDS solution and gently agitate by hand inversion after 3 h 
in the 37°C incubator. 

5. After a total of 18 h at 37°C, discard the solution and gently shake any oil drop- 
lets from the samples. 

6. Place bone samples in distilled water, invert several times, and discard the solu- 
tion. Repeat three times to remove any debris from the samples. 

7. Add fresh distilled water and leave for 3 h. Repeat this step once more for 3 h. 

8. Discard water and add acetone for 2 h. 

9. Replace acetone, invert the sample, and leave overnight. 

10. Next morning, replace acetone, invert, and leave all day. At the end of the day 
replace acetone again and leave for 2-3 d. 

1 1 . Remove acetone and shake off excess from the trabecular pores of the sample. 

1 2 . Add diethyl ether and leave all day. In the evening replace diethyl ether and leave overnight. 

13. Discard solution and gently shake the sample to remove any residue. 

14. Place samples in a 60°C oven for 8 h on a glass dish to evaporate any residue. 

15. Place samples under vacuum and leave overnight. 

16. Mount samples on aluminum stubs using silver dag adhesive. Allow to dry over- 
night in a desiccator. 

17. Coat samples with 20-nm platinum and examine in the scanning electron micro- 
scope at 10 kV to obtain a SE image. 

18. Alternatively, coat the samples thinly with carbon and obtain a BSE image at 20- 
30 kV. If both a SE and BSE image are required from the same specimen see 
Notes 6 and 7 and Fig. 2. 

3.2.2. Embedding of Anorganic Samples in LR White and Processing 
for BSE Imaging 

Resin embedding is a technique available in many EM laboratories and tech- 
nical advice on the methods available locally should be obtained. The general 
steps are given below. 

22/Aspden/311-320/F1 316 1 2/26/03, 10:48 AM 

I ■ I I I 


SEMofBone 317 

1. Prepare bone specimens as described in Subheading 3.2.1., steps 1-15. 

2. Embedding is done in a low-viscosity acrylic resin (LR White; see Note 8) under 
vacuum to ensure total impregnation. The resin is thermally cured without the 
need for a catalyst at a temperature of 55°C for 18-24 h. 

3. After curing, each specimen must be sectioned using a diamond saw (see Sub- 
heading 3.2., step 2) in a plane parallel with the surface to be examined. 

4. The cut face must then be polished; a roughness of <1 u,m is generally required 
for the highest quality work. This is done using a lapping machine using silicon 
carbide lapping plates and successively finer grades of diamond paste. 

5. A slab containing the prepared surface, approx 5 mm thick, is then cut off, 
mounted on an SEM stub with carbon tape and carbon coated (see Note 2). 

6. Examine the specimen using BSE at an accelerating voltage of 20-30 kV (see 
Note 7 and Fig. 3). 

4. Notes 

1. To avoid formation of calcium phosphate crystals, fixation in cacodylate buffer 
should be considered (see the chapter by Everts et al., this volume). However, at 
lower magnifications, we find this not to be a significant problem and routinely 
use the nontoxic phosphate buffers. 

2. Coating of the specimen is essential to make it electrically conductive. This can 
be done with gold-palladium, platinum, aluminum, or carbon. We routinely use 
platinum for SE imaging. A thin layer of carbon coating should be used for quan- 
titative analyses using the BSE mode, as the heavy metals interfere with the BSE 

3. A traditional means for removing the organic component of bone is to use bleach 
(5% sodium hypochlorite). This is very effective. We prefer to use the combina- 
tion of enzyme and detergent as it is less aggressive and the enzyme will target 
specifically the proteins while the detergent will help with their solubilization as 
well as that of most fats and carbohydrates. Collagen forming the structure of the 
bone will be protected by the mineralization and in any case is resistant to Pro- 
teinase K. 

4. For storage for longer than a few d, samples are wrapped in several layers of 
gauze or tissue soaked in phosphate-buffered saline (PBS); double-bagged in 
evacuated, sealed plastic bags; and frozen at lower than -20°C. Short-term stor- 
age can be done in a refrigerator in PBS. If surface mineral content is important, 
then storage is best in a saturated calcium phosphate buffered solution as slow 
dissolution of surface calcium phosphate has been observed. We have used a 
calcium phosphate buffered 0.15 M saline solution (containing 0.2 mM 
CaCl 2 .2H 2 0, 0.2 mM Na 2 HP0 4 , 0.01 mM Na 4 P 2 O 7 .10H 2 O and 0.4 g of sodium 
azide per liter as a bactericide/fungicide), as this has been shown to preserve the 
structure and composition of the bone (15). 

5. If freeze-fracturing is employed then the intention is normally to study the frac- 
ture surface. Clearly, any inspection of the surfaces that have been gripped is not 
advised and if this is done on cancellous bone then there is a distinct risk that 

22/Aspden/311-320/F1 317 I 2/26/03, 10:48 AM 

I ■ I I I 


Marshall et al. 


Fig. 3. BSE image of a resin-embedded bone slice. Note the clear distinction in 
brightness between different areas, indicative of differences in mineralization density. 



some crushing of the specimen could occur. If BSE is used to examine the frac- 
tured surface remember that surface topography contributes substantially to the BSE 
signal and only the "flatter" surfaces will give interpretable results (see Fig. 2). 
An accelerating voltage of 10 kV is routinely used. For BSE, an accelerating 
voltage of 20-30 kV is required. The carbon coating of the specimen for BSE 
should be thin to allow sufficient electrons to penetrate the specimen. If both and 
SE and BSE image from the same specimen are required, use carbon coating. 
Quantitative BSE imaging uses detection of electrons backscattered in the top 
(0.5 jim) surface of the specimen. Contrast in the BSE image depends on the 
mean atomic number. Bone contains organic material and mineral with calcium 
being the constituent with the highest atomic number. Calcium therefore deter- 
mines in largest part the intensity of the backscattered electrons. The BSE image 
appears brighter with increasing mineral density, which is related to bone turn- 
over rate, with older bone having a higher density than newer bone (1). The image 
obtained contains areas with different gray levels (Fig. 3). By carefully calibrat- 
ing the BSE signal with reference material of known atomic number, this gray 
scale can be used to obtain regional calcium concentrations. Different calibration 
procedures have been described (11,16) and the reader is referred to these origi- 
nal texts for analysis methods. 



2/26/03, 10:48 AM 

I ■ I I I 


SEMofBone 319 

8. For quantitative analysis LR White may be less appropriate as it has a high chloride 
content and this needs to be corrected for (9). Embedding in PMMA is a more 
common but more time consuming procedure. Addition of styrene to PMMA has 
been described to obtain resin that has improved stability under electron bombard- 
ment (1,16). This modification further increases specimen preparation time and 
because this is not a routine procedure in pathology laboratories, may only be prac- 
tical for research specimens. The choice of embedding material is therefore a trade 
off between what is best and what is practical. In most recent publications PMMA 
embedded material was examined. Generally, resin laboratories will be able to adapt 
their protocols for soft tissue to hard tissue requirements, essentially by increasing 
impregnation times. Very careful polymerization is also essential to avoid over- 
heating leading to generation of air bubbles in blocks leading to holes and cracks in 
sections. Such blocks are not suitable for BSE analysis, because of their uneven 
surface and careful selection of archive material is necessary. 

Acknowledg merits 

The authors are grateful to Dr. Helena Benford and Dr. Mike Rogers for 
providing Fig. 1. 


1. Boyde, A., Maconnachie, E., Reid, S. A., Delling, G., and Mundy, G. R. (1986) 
Scanning electron microscopy in bone pathology: review of methods, potential 
and applications. Scan. Electr. Microsc. IV, 1537-1554. 

2. Boyde, A. and Jones, S. J. (1996). Scanning electron microscopy of bone: Instru- 
ment, specimen and issues. Microsc. Res. Tech. 33, 92-120. 

3. Goldman, H. M., Kindsvater, J., and Bromage, T. G. (1999) Correlative light and 
backscattered electron microscopy of bone — Part I: specimen preparation meth- 
ods. Scanning 21, 40-42. 

4. Goldman, H. M., Blayvas, A., Boyde, A., Howell, P. G., Clement, J. G., and 
Bromage, T. G. (2000) Correlative light and backscattered electron microscopy 
of bone — Part II: automated image analysis. Scanning 22, 337-344. 

5. Abe, K., Kanno, T., and Schneider, G. B. (1983) Surface structure and osteoclasts 
of mouse parietal bones: a light and scanning electron microscopic study. Arch. 
Histol. Jpn. 46, 663-667. 

6. Chambers, T. J. and Fuller, K. (1985) Bone cells predispose bone surfaces to 
resorption by exposure of mineral to osteoclastic contact. /. Cell Sci 76, 155-165. 

7. Chambers, T. J., Revell, P. A., Fuller, K., and Athanasou, N. A. (1984) Resorp- 
tion of bone by isolated rabbit osteoclasts. /. Cell Sci. 66, 383-399. 

8. Jones S. J., Boyde, A., and Ali, N. N. (1984) The resorption of biological and non- 
biological substrates by cultured avian and mammalian osteoclasts. Anat. 
Embryol. 170, 247-256. 

9. Li, B., Marshall, D., Roe, M., and Aspden, R. M. (1999) The electron microscope 
appearance of the subchondral bone plate in the human femoral head in osteoar- 
thritis and osteoporosis. /. Anat. 195, 101-110. 

22/Aspden/311-320/F1 319 1 2/26/03, 10:48 AM 


I ■ I I I 

320 Marshall et al. 

10. Roscher, P, Fratzl, P., Klaushofer, K., and Rodan, G. (1997) Mineralisation of 
cancellous bone after alendronate and sodium fluoride treatment: a quantitative 
backscattered electron imaging study on minipig ribs. Bone 20, 393-397. 

11. Roschger, P., Fratzl, P., Eschberger, J., and Klaushofer, K. (1998) Validation of 
quantitative backscattered electron imaging for the measurement of mineral den- 
sity distribution in human bone biopsies. Bone 23, 319-326. 

12. Kingsmill, V. J. and Boyde, A. (1998) Mineralisation density of human mandibular 
bone: quantitative backscattered electron image analysis. /. Anat. 192, 245-256. 

13. Boyde, A., Travers, R., Glorieux, F. X., and Jones, S. J. (1999) The mineralization 
density of iliac crest bone from children with osteogenesis imperfecta. Calcif. 
Tissue Int. 64, 185-190. 

14. Friel, J. J. (1995). X Ray and Image Analysis in Electron Microscopy. Princeton 
Gamma Tech Inc., Princeton, NJ. 

15. Lees, S. (1988) Sonic velocity and the ultrastructure of mineralised tissues, in 
Calcified Tissues (Hukins, D. W. L., ed.), Macmillan, London, UK, pp. 121-152. 

16. Boyde, A., Jones, S. J., Aerssens, J., and Dequeker, J. (1995) Mineral density 
quantitation of the human cortical iliac crest by backscattered electron image 
analysis: variations with age, sex and degree of osteoarthritis. Bone 16, 619-627. 

22/Aspden/311-320/F1 320 1 2/26/03, 10:48 AM 


Bone Measurements by Peripheral Quantitative 
Computed Tomography in Rodents 

JCirg A. Gasser 

1. Introduction 

Quantitative computed tomography (QCT) is an established technique for 
the determination of bone mineral density (BMD) in the axial and appendicu- 
lar skeleton (I). QCT is unique amongst methods of bone mineral measure- 
ment in providing separate estimates of trabecular and cortical bone mineral 
density as a true volumetric mineral density value (g/cm 3 ) (2). In addition, 
QCT can measure geometric properties of cortical bone with great accuracy 
(3) and predict some mechanical properties with remarkable precision (4-6). 
Peripheral quantitative computed tomography (pQCT) is a special type of com- 
puted tomography in which scans of the appendicular skeleton are performed 
at a low radiation dosage. Bone and muscle development can be assessed 
noninvasively by pQCT at peripheral sites in studies of bone development, 
experimental models of bone loss, and in monitoring the effectiveness of thera- 
peutic interventions. In addition, pQCT can be used to assess excised bones ex 
vivo from virtually any skeletal site. 

The XCT 900 and XCT 960 series of pQCT scanners were developed at the 
University of Wiirzburg (7), and brought into commercial clinical and preclini- 
cal application by Stratec Medizintechnik GmbH, Germany. The present chap- 
ter focuses entirely on the two most frequently used pQCT scanners which 
were adapted for use in small rodents, the XCT 960A and the XCT 2000. All 
pQCT scanners of the XCT-series work according to the translation-rotation 
principle. The photons emitted by the X-ray tube are detected by 6 (XCT 960A) 
or 12 (XCT 2000) semiconductor detectors which have a near 100% efficacy 
for X-rays of around 38.5 keV. The attenuation coefficients at each point of the 
cross-sectional image are reconstructed from the projected data (filtered back- 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 




Table 1 

Technical Data of the XCT Series 

XCT 960 

XCT 960A 

XCT 960M 

XCT 2000 

Slice thickness 

2.2 mm 

1.2 mm 

1.0 mm 

0.55 mm 

High voltage 

47 kV 

49 kV 

35 kV 

45-49 kV 

X-Ray energy 





Matrix size 






590 Jim 

90-500 \im 

70-500 \im 

70-500 \im 

No. of projections 





Object size 

Up to 85 mm 

Up to 85 mm 

Up to 85 mm 

Up to 50 mm 

Scan length 

30 mm 

30 mm 

30 mm 

30 mm 

Radiation dose 

0.01 mSv/h 

<0.1 mSv/h 

<0.1 mSv/h 

<0.1 mSv/h 

The XCT960 scanner was originally developed for clinical use. The machine was later adapted 
for use in rats, dogs, and primates (XCT960A) and mice (XCT960M and XCT2000) by changing 
the collimator to decrease slice thickness and obtain better resolution. 

projection) (8). Daily calibration of the system using a hydroxyapatite contain- 
ing phantom allows one to calculate density values (mg/cm 3 ) from the attenu- 
ation coefficients. The XCT960A was the first commercially available 
dedicated animal pQCT scanner and in its technical aspects is closely related 
to the newer XCT Research SA. It was derived from the XCT 960 used in the 
clinic but provides a higher spatial resolution (100 ^im) and smaller slice thick- 
ness (1.2 mm) to account for the reduced size of bones of small animals such as 
mice, rats, dogs, primates, and sheep (Table 1). All the rat data presented in 
this chapter were measured on the XCT 960A. The XCT2000 scanner evolved 
from the dedicated mouse scanner XCT 960M, which offered higher spatial 
resolution (70 fxm), decreased slice thickness (0.55 mm), and better perfor- 
mance, and is closely related to the XCT Research M. The XCT2000 scanner 
was used for all the mouse data presented in this chapter. Most of the pQCT 
data reported in the literature is derived from measurements carried out in 
Sprague-Dawley or Wistar rats (2-6,9,10). However, the increasing use of 
transgenic and KO animals in skeletal biology, with the aim of associating a 
skeletal phenotype with loss of function of a gene, receptor, or protein or with 
its overexpression, has increased our need to develop similar methods for mice 
(11,12). Systems such as the XCT 2000 or XCT Research M fulfil these 
requirements. For general, more detailed information on the effects of ovariec- 
tomy (OVX) on the skeleton in rats the reader is referred to articles published 
elsewhere (13-15). 

pQCT in Small Animals 325 

2. Materials 

1. XCT 960A and XCT 2000 scanners. Stratec Medizintechnik GmbH, Germany. 

3. Methods 

3. 1. Number of Animals 

The number of rats or mice per group, as a general rule, should be a mini- 
mum of 10 or higher to guarantee statistical significance for primary outcome 
parameters derived from pQCT measurements. This recommendation is based 
on the coefficient of variation obtained on the XCT 960A (rats) and on the 
XCT2000 (rats and mice) (Table 2). The coefficient of variation in percent 
was calculated using the formula (CV = STDEV/Mean * 100). Instrument pre- 
cision (CV ; %) was determined by repeating 10 measurements in the proximal 
tibia ex vivo, without repositioning of the bone at a distance of 5 mm from the 
proximal end of the bone. In vivo precision was calculated from 10 measure- 
ments with (CV r %) or without repositioning of the limb CV U %). Very compa- 
rable values for CV r were obtained in rats and mice on the XCT2000. We 
recommend that users determine their own set of parameters, as these values 
are dependent on species, number of animals per group, age of animal, instru- 
ment type, scanning location, limb holder, and instrument setting. As a "rule of 
thumb" the difference between two measurements must be at least three times 
the value of the CV displayed in Table 2 for you to obtain statistically signifi- 
cant differences. 

3.2. Age of Animals 

Unless the focus of the study is on the evaluation of skeletal growth or fac- 
tors influencing it, rats aged 6 mo (or, better, 9 mo) should be used for your 
experiments. The use of younger animals will inevitably confound disease- or 
drug- induced effects with those of skeletal growth. Figure 1 clearly demon- 
strates that even though there is little change in cross-sectional bone area in 
aging 6- and 9-mo-old rats, periosteal modeling drifts are triggered by OVX in 
the younger, but not older animals. Fewer data are available from mice to make 
clear recommendations with regard to the ideal age range for studies. For 
C57BL/6J-mice, an age of 6 mo or older appears to be ideal. 

3.3. Baseline Measurements and Other Controls 

If you perform your measurements in vivo, the baseline value should be 
taken for every animal and changes induced by the disease or drug treatment 
measured against it. However, if you plan to carry out ex vivo measurements 
on excised bones such as lumbar vertebral bodies (which you cannot measure 
in vivo), or to carry out biomechanical tests or histomorphometric analyses of 



Table 2 

Coefficient of Variation (CV) Expressed in % Obtained in Rats on 

the XCT 960A and in Rats and Mice on the XCT2000 



i: XCT 960A 






cv v % 

CV r % 

cv r % 

CV r % 

Total bone mineral 








Total bone mineral 

mg/mm 3 






Total bone area 

mm 2 






Trabecular bone 







mineral content 

Trabecular bone 

mg/mm 3 






mineral density 
Trabecular bone 

mm 2 







Cortical bone 







mineral content 

Cortical bone 

mg/mm 3 






mineral density 
Cortical bone area 

mm 2 






Cortical thickness 







(circular ring 
Periosteal perimeter 














Bone strength index 

mm 4 *g/cm 3 






CV; was determined from 10 measurements of the proximal tibia metaphysis carried 
out ex vivo, while CV r and CV V were determined in vivo with and without repositioning of 
the animal, respectively. 

bone specimens, the addition of a baseline group (necropsy at baseline) is man- 
datory. Intervention studies may require additional control groups to collect 
bones prior to initiating active treatment. 

In longitudinal studies, we recommend a baseline measurement and alloca- 
tion of the animals to the various treatment groups based on their total cross- 
sectional BMD and their trabecular BMD. This should ensure homogeneity of 
variances between the groups. Assigning the animals to different groups on the 
basis of baseline measurements has advantages over random assignment. Most 

pQCT in Small Animals 


Total cross-sectional area 












- © - Sham (age 6 mo) 
• OVX (age 6 mo) 

- ^ - Sham (age 9 mo) 
— * — OVX (age 9 mo) 

Fig. 1. Increase in cross-sectional area in the tibia of 6- and 9-mo-old rats. Sham- 
operated rats still show some growth-related increase in cross-sectional area up to age 
7 mo. OVX-induces pronounced periosteal bone apposition in 6 but no longer in 9- 
mo-old rats. 

interventions will induce bone loss (OVX, immobilization) or bone gain (ana- 
bolic agents such as parathryroid hormone [PTH], growth hormone [GH], pros- 
taglandin E 2 [PGE 2 ]). Because the magnitude of the changes is in most cases 
proportional to the cancellous template, animals with different baseline BMD 
values will respond differently. Adjustment of groups to match for total cross- 
sectional BMD will also assure that animals are well matched in terms of bio- 
mechanical parameters. 

3.4. Transgenic Animals 

There are situations that do not allow a perfect protocol under consideration 
of Subheadings 3.1.-3.3. For example, transgenic or KO mice may show re- 
duced survival in the postnatal period and it may be impossible to study their 
mature skeletal phenotype or to conduct longitudinal studies. Many investiga- 
tors who interpret pQCT data derived from KO mice have not considered that 
the observed bone phenotype may derive from disturbances occurring during 
intrauterine skeletal morphogenesis, or changes in skeletal growth in the post 
natal period, concluding perhaps wrongly that the respective gene, receptor, or 
protein in question may represent an interesting target for the treatment of os- 

3.5. Selecting Regions of Interest 

For in vivo monitoring of bone mass, density, and cortical architecture, 
pQCT measurements are restricted to locations in the appendicular skeleton 
and tail vertebra. In the past, most of these measurements were carried out in 



Fig. 2. Choice of region of interest in skeletally mature rats and mice. Indicated is 
the distance in millimeters from the joint space for the distal femur metaphysis and the 
proximal tibia metaphysis. A pure "cortical" bone site can be easily assessed in the 
mid-diaphysis of the tibia (dotted line). 

hindlimbs (2-6,10). The proximal tibia metaphysis (PTM) has been the pre- 
ferred site for various reasons (Fig. 2). This site is rich in cancellous bone and 
reacts with the greatest magnitude of change to interventions such as OVX (2), 
immobilization, and bone anabolic therapies such as PTH. Another reason for 
choosing the PTM is the fact that it has been the most popular site for measur- 
ing structural and dynamic histomorphometric parameters. Serial scans carried 
out at various locations throughout the rat PTM showed that a section placed at 
a distance of 4.5-5 mm from the proximal end of the bone, a position in the 
secondary spongiosa, is most suited for skeletally mature animals (6 mo and 
older) (Fig. 3) (3). The ideal "corresponding" site in mice is located 2.5mm 
from the proximal end of the tibia (Fig. 2). Not surprisingly, the magnitude of 
cancellous bone loss in rats decreases when the section is placed further away 
from the knee joint, because of the decreasing amount of spongy bone found, 
especially in older animals. Another ideal site to study cancellous and cortical 
bone is the distal femur metaphysis (DFM) even though the local strain pat- 
tern, which generates large forces in the direction of the patella, limits the mag- 
nitude of cancellous bone loss at this site (2). For rats an ideal slice is placed at 
5 mm from the distal end of the femur (3 mm in mice). In contrast, the distal 
tibia metaphysis (DTM) is only of limited use for pQCT scanning. Its 
crosssection is small and separating cancellous and cortical bone is in most 
cases not possible in rats and mice. Because of the inaccessibility of lumbar 
vertebral bodies (LVB) for in vivo monitoring, it is tempting to carry out mea- 

pQCT in Small Animals 


Fig. 3. Scout-scan of knee joint and positioning of ROI at the 4.5-mm distance in 
the proximal tibia metaphysis of the rat. For better visibility, the inserted picture shows 
an X-ray taken at the same location. 

surements in tail vertebrae (TVB) instead, which are easy to position. Unfortu- 
nately the magnitude of bone loss in the TVB is considerably smaller than 
observed in PTM, DFM, and LVB, so that this site is only of limited use. The 
reasons for this "failure" are not clear but may include factors such as different 
cellularity (fatty marrow as opposed to red marrow), vascularity (metabolism), 
or protective local strain patterns. None of the scanner algorithms can separate 
cancellous and cortical bone in a perfect way. For this reason we recommend 
choosing a "pure" cortical bone site in addition to the PTM or DFM. One of the 
easiest sites to localize repeatedly with high precision in longitudinal studies is 
the mid-diaphysis (MDT) of the tibia. Bone length can be easily determined on 
a contact X-ray and the position calculated in growing as well as in skeletally 
mature rats and mice. 

3.6. Quality Assurance 

Before initiating your measurements in animals you must start the QA pro- 
cedure (quality assurance procedure) by scanning the QA phantom provided 
by the vendor. This procedure has to be carried out once every day before the 
machine will allow you to collect data. The routine starts by initiating a Scout 
View scan (SV scan) followed by the collection of data from the phantom on 

330 Gasser 

three different density values, all of which have to stay within the tolerated 
range of <1% of the true value. 

3.7. Anesthesia and Limb Positioning 

Accurate positioning of the limb is one of the crucially important factors to 
guarantee high-quality data in longitudinal studies. For this purpose we devel- 
oped a special holder that guarantees exact positioning and does not allow 
movement of the limb in any direction, as well as preventing its rotation. For 
the measurement, the animal is placed on a plastic tray in a lateral position. 
The mouth and nose of the animal are placed in a hole to which a tube is plugged 
on to deliver the inhalation anesthetic, namely 2.5% Isofluran (Forene®, Fig. 
4), which we deliver together with 0.8% oxygen and 0.8% air. Because the 
entire procedure of scanning normally requires < 6 min, this type of anesthesia 
is ideal. Rats and mice can be kept under anesthesia for prolonged periods if 
multiple slices are being assessed. Alternatively, anesthesia can be induced by 
intraperitoneal injection of 40 mg/kg Ketarom (1.2 mL of ketamine hydrochlo- 
ride + 0.8 mL of rompun 2% + 8 mL of NaCl). 

For proper positioning, the leg is placed into the tube at the other end of the 
holder with the foot sticking out (Fig. 5). The conical plug is then placed around 
the limb to hold the tibial muscle with the iron rod inserted into the lateral slit, 
thus preventing any rotation of the limb. Finally, the foot is kept from sliding 
back into the tube by placing the foot-holder around the ankle and securing it 
with adhesive tape (Fig. 5). The conical plug is available in three lengths and 
the right size is chosen by the operator depending on the size of the animal. 
This holder guarantees next to identical placement of the leg from one mea- 
surement to the next. The animal is now ready for the scout scan. Always give 
each animal a unique name (code) and keep this designation for all later time 
points. As soon as the code is typed in a second time, the software will recog- 
nize that it should compare the results of the upcoming SV scan with an earlier 
measurement from the same animal. This software-based automatic detection 
procedure is very able to place your measurement at the same location of any 
previous scan. 

3.8. Setting up the Scanner 

The numbers of possibilities for setting your instrument are virtually end- 
less and it is beyond the scope of this guide to discuss all of them. We were 
operating software version 5.40 on both scanners. In Table 3 you find our 
proposal for instrument setting for rats on the XCT960A and the XCT2000, as 
well as the settings for mice on the XCT2000. In our experience they work 
very well and give robust results of high quality for both cortical and cancel- 
lous parameters. We have chosen to use the same threshold based contour find- 

pQCT in Small Animals 


Fig. 4. Animals are placed into a lateral position under inhalation anesthesia with 
2.5% isoflurane (Forene®). Suction around the nose prevents spillage of fumes to the 
surrounding air. The overview shows the system with a special "flow through cham- 
ber" in which the next animal can be anesthetised while the limb of another rat is 

Legholder /"^ 


— ^m^ 

0t ^ 

Conical Plug Footholder 


Fig. 5. Detailed view of the leg holder consisting of a tube, a conically shaped plug 
with a rod, and the foot holder. The function of the conical plug is to hold the tibial 
muscle preventing its rotation. The foot holder is placed around the ankle preventing 
its withdrawal into the tube. 

ing for detection of both, the periosteal and the endocortical surface 
(CONTMODE 1, CORTMODE 2, PEELMODE 2) (Fig. 6). Some investiga- 
tors prefer to define the cancellous bone area as a fixed value of 45% of the 
total bone area. This mode of analysis may have advantages in larger animals 
such as dogs, primates, and humans, where the cortical compartment is more 
difficult to separate from cancellous bone by threshold-based algorithms. For 
rodents we clearly prefer our threshold-based separation procedure as it gives 



Table 3 

Instrument Setting for Measurements in the Proximal Tibia Metaphysis 

with the XCT 960A (Rats) and the XCT2000 (Rats and Mice) 

XCT960A (rats) 

XCT2000 (rats) 

XCT2000 (mice) 


1.2 mm 

1.0 mm 

0.5 mm 


Voxel size 

0.197 x 0.197x1.2 mm 

0.2 x 0.2 x 1 mm 

0.1 x 0.1 x 0.5 mm 

Scan speed: 


20 mm/s 

20 mm/s 

10 mm/s 

Final scan 

1 mm/s 

10 mm/s 

3 mm/s 














730 mg/cm 3 

730 mg/cm 3 

400 mg/cm 3 


730 mg/cm 3 

730 mg/cm 3 

400 mg/cm 3 

Threshold determines the outer contour in the CONTMODE part of scan-analysis while 
Threshold2 defines the cortical and cancellous bone compartment in CORTMODE. 

Fig. 6. Actual scan through the PTM at 4.5 mm from the knee joint taken on the 
XCT960A in a 9-mo-old rat. In most cases the software is able to define the ROI 
automatically on which the scan analysis is run. The dotted line delineates the contour 
of the muscle. Using a lower threshold, muscle area can be determined in all sections. 

pQCT in Small Animals 333 

much more accurate information on the changes in cortical thickness and 
endocortical perimeter. 

Our preferred setting for scan analysis is defined in the following subheadings. 

3.8.1. CONTMODE 1 

This parameter determines the separation of soft tissue from bone for delinea- 
tion of the outer contour (periosteal perimeter). The threshold attenuation coeffi- 
cient can be freely chosen by the operator and should be set to 610-730 mg/cm 3 
in rats and 350-400 mg/cm 3 in mice. 

3.8.2. CORTMODE2 

This algorithm separates trabecular from cortical bone at the endocortical 
surface. All voxels with a lower attenuation coefficient than the selected thresh- 
old are eliminated and counted as being part of the cancellous bone compart- 
ment. The voxels are then "proofed" by a 3 x 3 filter to ensure continuity. We 
recommend using the same threshold to define the inner contour of cortical 
bone as was used for detection of the outer contour (contour mode 1; 610-730 
mg/cm 3 in rats and 350-400 mg/cm 3 in mice). 

You may find it worthwhile to experiment with the instrument settings and 
find those which suit your purpose best (see Subheading 3.10.). 

3.9. Data Analysis 

The analytical part of the software generates a great number of parameters 
but luckily you do not require all of them. We find the 13 parameters listed in 
Table 1 are sufficient for adequate interpretation of your data. In the following 
we will attempt to describe such an analytical process on a true data set. The 
12-wk study was performed in 9-mo-old virgin Wistar rats. Ten animals per 
group were ovariectomized and treated daily p.o. with 0.3 mg/kg of 17a- 
ethinylestradiol dissolved in corn oil. Other groups of animals was injected 
twice weekly subcutaneously with 10 \ig/kg alendronate or vehicle (corn oil). 
A sham-operated group was also measured to determine age-related changes 
with pQCT measurements being carried out in all groups at baseline, 4, 8, and 
12 wk. Although it is possible to analyze absolute data in well-matched groups 
of animals, we recommend that you calculate the percent change for each pa- 
rameter and time point of all animals from their own baseline value. The rea- 
sons for this is simple: You will find it virtually impossible to adjust the values 
of all treatment groups for all 13 parameters that you want to evaluate when 
allocating the animals to their groups at baseline. When looking at your graphs 
containing the data, always keep in mind the "rule of thumb" mentioned ear- 
lier. For any "effect" to be real, the percent change should be at least three 
times bigger than the respective CV r value displayed in Table 2. 



Table 4 

Definition of Most Relevant Parameters Required for Data Interpretation 




TOT_CNT Total bone mineral content 

TOT_DEN Volumetric total bone mineral density 

TOT_A Total bone mineral area 

CRT_CNT Cortical bone mineral content 

CRT_DEN Volumetric cortical bone mineral density 

CRT_A Cortical bone mineral area 

TRAB_CNT Cancellous bone mineral content 

TRAB_DEN Volumetric cancellous bone mineral density 

TRAB_A Cancellous bone mineral area 

PERI_C Periosteal circumference 

ENDO_C Endocortical circumference 

CRT_THK Mean cortical thickness ring model 

xBSI Axial bone strength Index (bending strength) 

pBSI Polar bone strength index (torsional strength) 


mg/cm 3 

mm 2 


mg/cm 3 

mm 2 


mg/cm 3 

mm 2 




i 4 * 


g/cm 3 
g/cm 3 

Column 1 lists the parameters as denominated by the software and their unit is listed in col- 
umn 3. Only the most relevant parameters required for interpretation of the data are listed. 

We always start our analytical process by looking at the periosteal perimeter 
(PERI_C) or the total bone area (TOT_A) (Table 4). Both parameters describe 
the same feature of bone size. However, TOT_A, being an area measurement, 
shows a greater magnitude of change. Only an increase of 2% or greater in 
PERI_C or > 3% in TOT_A gives you an indication of a disease or drug-in- 
duced effect. Such changes are easily obtained during early skeletal growth up 
to the age of 6 mo in rats. Anabolic agents such as GH (9,16,17), vitamin D 
(18), and high dose PTH can induce them. In our example (Fig. 7A), none of 
the changes measured in PERI_C exceeded the 2% magnitude indicating that 
observed variations are within the precision of the measurement and therefore 
not real effects. 

The next parameter we consider is the total cross-sectional BMC 
(TOT_CNT), which indicates whether there is an absolute gain or loss in bone 
mass caused by the intervention. Together with the previous parameter 
(PERI_C), we can learn whether this increase is happening entirely at the 
endosteal envelope (i.e., in the absence of changes in PERI_C) or connected to 
an increase in bone size (increase in PERI_C). In our simple example there was 
no change in bone size; therefore TOT_CNT gives the same information as the 
volumetric cross-sectional BMD (TOT_DEN) displayed in Fig. 7B. Our "rule 
of thumb" tells us that the sham OP and the alendronate group are stable and 

pQCT in Small Animals 


Periosteal Perimeter (PERI_C) 










— *K— Ethinylestradiol 
O Alendronate 

u ™~ "^ " s 



Volumetric Cross-Sectional BMD (TOT_DEN) 



— *— Ethinylestradiol 

— S— Alendronate 

4 8 12 


Mean Cortical Thickness (CRT_THK) 

• OVX 
O Alendronate 


Fig. 7. (A) Change in periosteal perimeter (% from baseline) after OVX. (B) Change 
in volumetric cross-sectional BMD (% from baseline) after OVX. (C) Change in mean 
cortical thickness (% from baseline) after OVX. 

not different from each other over time. However, the curve for the 
ethinylestradiol-treated rats drops off by more than three times the CV r found 



Endocortical Perimeter (ENDO_C) 


■ OVX 

O Alendronate 


Volumetric Cancellous BMD (TRAB_DEN) 


— *K— Ethinylestradiol 
O Alendronate 

Volumetric Cortical BMD (CRT_DEN) 



-96- Ethinylestradiol 


Fig. 7. (D) Change in endocortical perimeter (% from baseline) after OVX. (E) 
Change in volumetric cancellous BMD (% from baseline) after OVX. (F) Change in 
volumetric cortical BMD (% from baseline) after OVX. 

pQCT in Small Animals 337 

in Table 2 (>0.8%), indicating that treatment is not 100% effective in prevent- 
ing bone loss. OVX rats show highly significant changes in TOT_DEN already 
at 4 wk and this parameter can already give meaningful information at a 2-wk 
measurement point. In the absence of any change in bone size we can now also 
conclude that bone loss must have occurred entirely on the endosteal envelope 
(cancellous or endocortical bone resorption). Our analytical process should 
therefore try to distinguish between these two possibilities. 

The best parameters available to address this question is to look for changes 
in mean cortical thickness (CRT_THK, Fig. 7C), and the threshold delineated 
endocortical perimeter (ENDO_C, Fig. 7D) or, alternatively, the cancellous 
bone area (TRAB_A). In our example, both the sham OP and alendronate- 
treated animals are stable over time in these three parameters. The 
ethinylestradiol-treated animals show changes that just about exceed three 
times the CV r value for this parameter displayed in Table 2 (>2.6%). This 
indicates that some cortical thinning through endocortical bone resorption has 
occurred. Apparently, this resorptive process is very fast and significant corti- 
cal thinning is observed in vehicle treated OVX rats as early as 4 wk after the 
operation (Fig. 7C). A decrease in CRT_THK in the absence of changes in 
bone size (PERI_C) indicates that all of the changes in CRT_THK are the re- 
sult of endocortical bone resorption. Treatments that are known to increase 
cortical thickness through endocortical bone apposition are PTH (2,19,20) and 
PGE 2 (21,22). The magnitude of these changes (decrease in ENDO_C) is large 
and can easily be monitored after 4 wk or longer. 

Next we would like to interpret the results on the trabecular compartment. 
Remember that we were using a threshold-based algorithm to separate cancel- 
lous from cortical bone. Also, our analysis shows that the marrow cavity is 
expanding through a strong endocortical resorption process. This ongoing 
endocortical erosion obviously lowers volumetric cortical density at the inter- 
face to such a degree, that it falls below the chosen threshold of 690 mg/cm 3 
and as a consequence is counted as cancellous bone. Because of the constant 
increase in TRAB_A, just looking at cancellous BMC (TRAB_CNT) may give 
the misleading impression of an increase in cancellous bone mass even though 
rapid removal of cancellous bone structures is ongoing. We therefore recom- 
mend using the area-corrected volumetric cancellous BMD (TRAB_DEN, Fig. 
7E) instead, which accounts for the expanding marrow cavity. This parameter 
clearly shows that cancellous bone is lost at even a greater rate than is observed 
in cortical thinning (CRT_THK, Fig. 7C). Again, alendronate appears to be 
fully protective while cancellous bone loss in ethinylestradiol-treated animals 
exceeds the threefold CV r displayed in Table 2, indicating small but signifi- 
cant bone loss from this compartment. 

338 Gasser 

Some of the known agents acting to increase cancellous bone in rats are 
PTH (23,24), fibroblast Growth factor (FGF) (25), PGE 2 (21,22,26), vitamin D 
analogues (27), and GH. 

Volumetric cortical BMD (CRT_DEN) (Fig. 7F) is the only parameter, 
which indicates an intrinsic material property of bone. In contrast to TOT_DEN 
and TRAB_DEN, which in fact are nothing but projected area density mea- 
surements similar to those obtained by DXA, CRT_DEN is a true volumetric 
measurement of mineral density. Under most circumstances, this parameter 
changes very little over time but it may increase in long-term studies with 
bisphosphonates to indicate increased matrix mineralization (stiffer bone). 
Conversely, this parameter may decrease at least transiently during long-term 
treatment with bone anabolic agents such as PTH, because the newly formed 
endocortical bone is less densely mineralised and because secondary mineral- 
ization takes some time to catch up (28). Volumetric CRT_DEN may also de- 
crease in cases in which bone mineralisation is disturbed (vitamin D deficiency, 
osteogenesis imperfecta) (29) or in the presence of increased Haversian re- 
modeling (only exceptionally in rodents). In our example there is indeed little 
change over time in CRT_DEN except for a trend to an increase in the 
alendronate and OVX group which, perhaps, could be indicative of increased 
matrix mineral content in bisphosphonate-treated animals. 

In summary, our example shows rapid cancellous and endocortical bone loss, 
the latter resulting in cortical thinning. Cortical thinning is not counteracted by 
activation of a compensatory periosteal bone formation drift in these skeletally 
mature animals, in contrast to what would be seen in younger rats (see Fig. 1). 
The magnitude of bone loss is greater in the cancellous than the cortical bone 
compartment. None of the interventions has significant effects on matrix 
mineralisation as concluded from cortical BMD measurements but there is at 
least a trend toward an increase in alendronate-treated rats. It may be worth- 
while to follow up such a trend with more sophisticated methods with better 
discriminatory power for matrix mineralisation such as backscattered electron 
imaging (30,31). 

3. 10. Tips and Tricks on Instrument Setting 

1. The settings you chose (especially the thresholds) vary with age of the animals. 
Younger animals often require lower thresholds than those displayed in Table 1. 
The same may be true for transgenic and KO mice with a severely osteopenic 
phenotype, animals that undergo heavy Haversian remodelling, or in situations 
in which mineralization is disturbed (vitamin D deficiency, osteogenesis 

2. You should choose the highest possible threshold for contour finding to mini- 
mize the partial volume effect. The ideal threshold is 730 mg/cm 3 , which corre- 

pQCT in Small Animals 339 

sponds to an attenuation coefficient of 0.94 cm -1 . This high threshold only really 
works for rats 9 mo of age and older and may fail even in this age group at later 
measurement points, if the intervention results in substantial cortical bone loss. 

3. If your cortical ring does not appear to be closed when running your chosen algo- 
rithm in the analytical part, your threshold is too high and you have to lower it, as 
the parameters describing cortical architecture would all be wrong! 

4. Within one experiment you must choose the same peeling algorithms including the 
thresholds for each animal and each measurement point! If your chosen setup fails 
during later time points in your experiment you must go back and reanalyze all 
your earlier data points of the entire experiment using lower thresholds therefore. 

5. We recommend to run the scan analysis loop always twice with two different 
thresholds (an ideal and a lower one). The two thresholds we routinely chose are 
690 and 590 mg/cm 3 , respectively, in skeletally mature rats, and 400 and 350 mg/ 
cm 3 for skeletally mature mice. If your experimental conditions result in strong 
bone loss and your initially chosen "higher" threshold fails during later data ac- 
quisitions, you can choose to work with the data of the "lower," nonoptimal sec- 
ond threshold without having to reanalyze all your earlier time points. 


1. Guglielmi, G., Gliier, C. C, Majumdar, S., Blunt, B. A., and Genant, H. K. (1995) 
Current methods and advances in bone densitometry. Eur. Radiol. 5, 129-139. 

2. Gasser, J. A. (1997) Quantitative assessment of bone mass and geometry by pQCT 
in rats in vivo and site specificity of changes at different skeletal sites. /. Jpn. Soc. 
Bone Morphom. 7, 107-114. 

3. Gasser, J. A. (1995) Assessing bone quantity by pQCT. Bone 17, S145-S154. 

4. Ferretti, J. L., Capozza, R. F., and Zanchetta, J. R. (1995) Mechanical validation 
of a tomographic (pQCT) index for non-invasive estimation of rat femur bending 
strength. Bone 17, S145-S162. 

5. Ferretti, J. L., Capozza, R. F., and Zanchetta, J. R. (1995) Mechanical validation of 
a non-invasive (pQCT) index of bending strength in rat femurs. Bone 18, 97-102. 

6. Ferretti, J. L. (1997) Non-invasive assessment of bone architecture and biome- 
chanical properties in animals and humans employing pQCT technology. /. Jpn. 
Soc. Bone Morphom. 7, 115-125. 

7. Schneider, P. and Borner, W. (1991) Peripheral quantitative computed tomogra- 
phy for bone mineral measurements using a new special QCT-scanner: methodol- 
ogy, normal values, comparison with manifest osteoporosis. Fortschr. Rontgenstr. 
154, 292-299. 

8. Hermann, G.T. (1980) Image Reconstruction from Projections: The Fundamen- 
tals of Computerized Tomography. Academic Press, Orlando. 

9. Banu, M. J., Orhii, P. B., Mejia, W., et al. (1999) Analysis of the effects of growth 
hormone, voluntary exercise and food restriction on diaphyseal bone in female 
F344 rats. Bone 25, 469-480. 

10. Breen, S. A., Millest, A. J., Loveday, B. E., Johnstone, D., and Waterton, J. C. 
(1996) Regional analysis of bone mineral density in the distal femur and proximal 

340 Gasser 

tibia using peripheral computed tomography in the rat in vivo. Calcif. Tissue Int. 

11. Beamer, W. G., Donahue, L. R., Rosen, C. J., and Baylink, D. J. (1996) Genetic 
variability in adult bone density among inbred strains of mice. Bone 18, 397-403. 

12. Graichen, H., Lochmiiller, E. M., Wolf, E., et al. (1998) A non-destructive tech- 
nique for a 3-D microstructural phenotypic characterisation of bones in geneti- 
cally altered mice: preliminary data in growth hormone transgenic animals and 
normal controls. Anat. Embryol. 199, 239-248. 

13. Wronski, T. J., Dann, L. M., Scott, K. S., and Cintron, M. (1989) Long-term effects 
of ovariectomy and aging on the rat skeleton. Calcif. Tissue Int. 45, 360-366. 

14. Yamazaki, I. and Yamaguchi, H. (1989) Characteristics of an ovariectomized 
osteopenic rat model. /. Bone Miner. Res. 4, 13-22 

15. Kalu, D. N. (1991) The ovariectomized rat model of postmenopausal bone loss. 
Bone Miner. 15, 175-192. 

16. Andreassen, T. T., Jorgensen, P. H., Flyvbjerg, A., Orskov, A., and Oxlund, H. 
(1995) Growth hormone stimulates bone formation and strength of cortical bone 
in aged rats. /. Bone Miner. Res. 10, 1057-1067. 

17. Andreassen, T. T. and Oxlund, H. (2000) The influence of combined parathyroid 
hormone and growth hormone treatment on cortical bone in aged ovariectomized 
rats. /. Bone Miner. Res. 15, 2266-2275. 

18. Weber, K., Goldberg, M., Stangassinger, M., and Erben, R. G. (2001) la- 
hydroxy vitamin D 2 is less toxic but not bone selective relative to la- 
hydroxy vitamin D 3 in ovariectomized rats. /. Bone Miner. Res. 16, 639-651. 

19. Ejersted, C., Andreassen, T. T., Oxlund, H., et al. (1993) Human parathyroid hor- 
mone (1-34) and (1-84) increase the mechanical strength and thickness of cortical 
bone in rats. /. Bone Miner. Res. 8, 1097-1 101. 

20. Ejersted, C., Andreassen, T. T., Nilsson, M. H., and Oxlund, H. (1994) Human 
parathyroid hormone (1-34) increases bone formation and strength of cortical bone 
in aged rats. Eur. J. Endocrinol. 130, 201-207. 

21. Jee, W. S. S., Mori, S., Li, X. J., and Chan, S. (1990) Prostaglandin E2 enhances 
cortical bone mass and activates intracortical bone remodeling in intact and ova- 
riectomized female rats. Bone 11, 253-266. 

22. Jee, W. S. S., Ke, H. Z., and Li, X. J. (1991) Long-term anabolic effects of pros- 
taglandin-E2 on tibial diaphyseal bone in male rats. Bone Miner. 15, 33-55. 

23. Gunness-Hey, M. and Hock, J. M. (1984) Increased trabecular bone mass in rats 
treated with synthetic parathyroid hormone. Metab. Bone Rel. Dis. 5, 177-181. 

24. Gunness-Hey, M. and Hock, J. M. (1993) Anabolic effect of parathyroid hormone 
on cancellous and cortical bone histology. Bone 14, 277-281. 

25. Pun, S., Dearden, R. L., Ratkus, A. M., Liang, H., and Wronski, T. J. (2001) 
Decreased bone anabolic effect of basic fibroblast growth factor at fatty marrow 
sites in ovariectomized rats. Bone 28, 220-226. 

26. Mori, S., Jee, W. S. S., and Li, X. J. (1992) Production of new trabecular bone in 
osteopenic ovariectomized rats by prostaglandin E2. Calcif. Tissue Int. 50, 80-87. 

pQCT in Small Animals 34 1 

27. Erben, R.G., Bromm, S., and Stangassinger, M. (1998) Therapeutic efficacy of 
la,25-hydroxyvitamin D3 and calcium in osteopenic ovariectomized rats: evi- 
dence for a direct anabolic effect of la,25-hydroxyvitamin D3 on bone. Endocri- 
nology 139,4319-4328. 

28. Kneissel, M., Boyde, A., and Gasser, J. A. (2001) Bone tissue and its mineraliza- 
tion in aged estrogen-depleted rats after long-term intermittent treatment with 
parathyroid hormone (PTH) analog SDZ PTS 893 or human PTH(l-34). Bone 28, 

29. Boyde, A., Travers, R., Glorieux, F.H., and Jones, S.J. (1999) The mineralisation 
density of iliac crest bone from children with osteogenesis imperfecta. Calcif. 
Tissue Int. 64, 185-190. 

30. Boyde, A., Jones, S. J., Aerssens, J. and Dequeker, J. (1995) Mineral density quan- 
tification of the human cortical illiac crest by backscattered electron image analy- 
sis: Variations with age, sex, and degree of osteoarthritis. Bone 16, 619-627. 

31. Roschger, P., Plenk, H., Jr., Klaushofer, K., and Eschberger, J. (1995) A new 
scanning electron microscopy approach for the quantification of bone mineral 
distribution: Backscattered electron image grey levels correlated to calcium 
Kalpha-line intensities. Scan Microsc. 9, 75-88. 

I ■ I I I 



Studies of Local Bone Remodeling 

The Calvarial Injection Assay 

Robert J. van 't Hot, Claire E. Clarkin, and Kenneth J. Armour 

1 . Introduction 

There are several assays available to study the effects of cytokines, drugs, 
and hormones on bone cells in vitro. However, as the complex interactions 
between cells are disrupted, these in vitro assays do not always reflect what 
happens in vivo. The calvarial injection method, originally described by Boyce 
et al. (1), is valuable for studying the effects of substances on bone metabolism in 
vivo. In this assay, the substance to be tested is injected subcutaneously over the 
calvarium of a mouse. At the end of the assay the animal is euthanized, and the 
calvarium dissected and analyzed by microscopy. Although the assay was origi- 
nally used to study effects of cytokines on osteoclast formation and activity (1 ,2), 
it has also been used to study the effects of drugs on bone formation (3). 

2. Materials 

2. 1. Injection 

1. Recombinant murine interleukin-la (IL-la) (5 mg/mL; CN Biosciences [UK] 
Ltd., Nottingham, UK). 

2. Hamilton syringe (Luer-lock type; Anachem Ltd, Luton, UK). 

2.2. Tissue Processing 

1. Histocryl (glycol methacrylate, TAAB). 

2. Resin mix: Add 1.5 g of catalyst (benzoyl peroxide, comes with the histocryl) to 
100 mL of histocryl; keep at 4°C. 

3. Accelerator mix: 5 mL of polyethylene glycol 4000 (PEG 400), 5 mL dibutyl 
pthalate, and 240 jjL of Histocryl accelerator. 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


24/van 't Hof/343-352/F1 345 1 2/26/03, 10:48 AM 

I ■ I I I 


346 van 't Hot et al. 

4. Embedding mix: Add 175 jjL of accelerator mix to 1 mL of resin mix (at 4°C) 
and use immediately. 

2.3. Tartrate-Resistant Acid Phosphatase (TRAP)Zvon Kossa/ 
Light Green Stain 

1. 1.5% (w/v) Silver nitrate in dH 2 0. 

2. 0.1% (w/v) Hydroquinone. 

3. 1% (w/v) Light green in dH 2 0. 

4. All the reagents for the TRAP stain are described in Chapter 1 1 by van 't Hof, 
this volume. 

2.4. Goldner's Trichrome Stain 

1. Weigert's hematoxylin: 

a. Solution A: Dissolve 1 g hematoxylin in 1 000 mL of absolute alcohol. Ripen 
for at least 4 wk before use. 

b. Solution B: Dissolve 11.6 g of ferric chloride (hydrated) in 1000 mL of dis- 
tilled water and add 10 mL of 2% HC1. 

Immediately before use mix equal parts of A and B. Do not keep working solu- 
tion premade. 

2. Ponceau de Xylidine-acid fuchsin: 1.5 g of Ponceau de Xylidine, 0.5 g of acid 
fuchsin, 2 mL of acetic acid (concentrated), and 98 mL of distilled water. 

3. Azophloxine (working solution): 0.5 g of azophloxine, 0.6 mL of acetic acid 
(concentrated), and 99.4 mL of distilled water. 

4. Ponceau de Xylidine-acid fuchsin-azophloxine (working solution): 12 mL of 
Ponceau de Xylidine-acid fuchsin, 8 mL of azophloxine, 80 mL 0.2% acetic acid; 
reuse the working solution. 

5. Phosphomolybdic acid-orange G: 6 g of phosphomolybdic acid, 4 g of Orange 
G, and 1000 mL of distilled water. 

6. Light green: 2 g of light green, 2 mL of acetic acid (concentrated), and 1000 mL 
of distilled water. 

3. Methods 

3. 1. Injection Protocol (Resorption) 

1 . Inject the mice over the calvarial bones with 1 \xL of recombinant murine IL- 1 a 
(5 mg/mL) or vehicle (sterile saline) using a 50-fxL Hamilton syringe. Perform 
injections three times per day for 3 consecutive days (see Notes 1-3). 

2. Euthanize the mice 4 d after the last injection. 

3. Dissect out the calvarial bones and fix in 4% buffered formalin-saline, pH 7.4, for 1 h. 

4. Rinse the calvaria in PBS and store in 70% alcohol. 

5. Embed the undecalcified calvarial bones in glycol-methacrylate (GMA, see 
Notes 4 and 5) and cut 3-fxm sections on a microtome (Jung, Heidelberg, Ger- 
many) using a glass knife (see Subheading 3.2. for embedding procedure). 

6. Stain sections with von Kossa and TRAP, followed by counterstaining with light 
green. Alternatively, especially when one is interested in effects on osteoblasts, 

24/van 't Hof/343-352/F1 346 1 2/26/03, 10:48 AM 

I ■ I I I 


Calvarial Injection Assay 347 

the sections can be stained with Goldner's trichrome (see Subheadings 3.3. and 
3.4. for staining protocols). 

3.2. Tissue Processing 

Cut out a strip of calvarial tissue from the center of the calvarium as illus- 
trated in Fig. 1. The following steps are most easily performed using a tissue 
processor, but can also be performed manually. All the steps are performed at 
4°C (see Note 6). 

1. Transfer tissue strips to 96% ethanol for 1 h. 

2. 1 h in 100% ethanol. 

3. 1 h in a 1:1 mix of 100% ethanol and resin mix. 

4. 1 h in resin mix. 

5. 72 h in resin mix. 

6. Transfer tissue to a mold placed in a crushed-ice slush, fill with embedding mix, 
seal with a stub, and leave to polymerize for 1 h. 

3.3. TRAP/von Kossa/Light Green Staining of Mouse Calvariae 

This method stains osteoclasts bright red, the mineralized bone black, and 
the remaining tissue green (Fig. 2). The von Kossa stain should be performed 
first because the TRAP staining solution (which is acidic) will dissolve much 
ofthe mineral from the section resulting in an unsatisfactory von Kossa stain. 

1. Immerse sections in 1.5% silver nitrate (made up when required and filtered just 
before use) for 40 s. 

2. Wash three times in water. 

3. Develop the stain in 0.1% hydroquinone for 25-30 s (maximum). Check using a 
microscope at this point; mineralized bone should be black, not brown. If the 
bone looks brown, rinse in water and repeat the procedure. 

4. Thoroughly rinse sections in running tap water for 10 min. Hydroquinone inhib- 
its TRAP staining, and this step ensures that all the hydroquinone is washed off. 

5. Perform the TRAP stain as described for cocultures in Chapter 1 1 by van 't Hof, 
this volume. Slides should be lying flat in plastic slide boxes with damp tissue 
lining the bottom. Boxes should then be covered to avoid drying of the staining 

6. Incubate at 37°C for 1.5 h. (Check staining after 1 h.) 

7. Rinse off the TRAP staining solution with dH 2 0. 

8. Counterstain with 1% light green for 30-60 s. Wash off with dH 2 0. 

9. Air-dry. 

10. Mount with aqueous mounting medium (e.g., Apathy's). 

11. Store in cardboard slide trays and cover to prevent fading. 

3.4. Goldner's Trichrome 

This stain results in bright green stained calcified bone and good contrast of 
the cells. Although the osteoclasts do not stand out as well as with the TRAP 

24/van 't Hof/343-352/F1 347 I 2/26/03, 10:48 AM 

I ■ I I I 


van 't Hot et al. 


Cut out strip of calvaria as indicated 

Embed, section and stain 

Analyse by microscopy 

Fig. 1. A strip of tissue is cut out of the fixed calvarium and embedded for process- 
ing as indicated in this figure. 

stain, this stain allows easy identification of osteoblasts. It is essential not to let 
the sections dry at any time during the staining protocol, as this leads to cracks 
in the mineralized bone. 

1. Keep the sections in distilled water for at least 1 h (to prevent bubbling below the 
section; if this still persists keep the slides in water for a longer time). 

2. Stain sections in Weigert's hematoxylin for 20 min (see Note 7). 

3. Wash in water. 

4. Differentiate with 0.5% acid alcohol. 

5. Wash in water for 20 min. 

6. Stain sections in Ponceau-acid fuchsin-azophloxine for 5 min. 

7. Rinse in 1% acetic acid for 10 s. 

8. Stain sections in phosphomolybdic acid-Orange G for 20 min. 

9. Repeat step 7. 

10. Stain sections in 0.2% light green for 5 min. 

11. Rinse in water. 

12. Blot dry. 

13. Rinse in 100% alcohol. 

14. Immerse the sections in xylene. 

15. Wipe off xylene around section before mounting in DePeX. 

This method stains cell nuclei blue/black, mineralized bone/muscle green 
and osteoid/collagen red (Fig. 3). 

24/van 't Hof/343-352/F1 


2/26/03, 10:48 AM 


Calvarial Injection Assay 



Fig. 2. Calvarial sections from neonatal mice treated with saline (A) or IL-la (B) 
Stained by TRAP and von Kossa. Unlike the control section (A), numerous TRAP- 
stained osteoclasts (arrows) are visible on the bone surface in (B) and extensive bone 
resorption is evident. 

Fig. 3. Calvarial section, stained with Goldner's trichrome, from a neonatal mouse 
treated with bone morphologenic protein (BMP-2). Large activated osteoblasts and 
osteoid/collagen are clearly visible above the bone surface. (Reduced from original 
magnification, x200.) 

24/van 't Hof/343-352/F1 


2/26/03, 10:48 AM 

I ■ I I I 


350 van 't Hot et al. 

3.5. Analysis of Results 

Although many qualitative conclusions can be drawn about the effects of 
test substances by simple microscopical observation, we usually perform a 
quantitative analysis using computer-assisted histomorphometry. Parameters 
of interest are numbers of osteoclasts and osteoblasts per bone surface, miner- 
alized bone width, and bone formation and resorption surfaces. We use soft- 
ware developed using the Aphelion ActiveX image analysis toolkit from 
ADCIS (ADCIS SA, Herouville-Saint-Clair, France) and a Zeiss Axioskop 
microscope fitted with a color camera. The program prompts the user to select 
and focus a field, captures the image, and identifies the part of the image that 
contains tissue. The bone is identified using color thresholding and the bone 
surface, volume, and width are calculated (see Note 8). Then the resorption 
and formation surfaces are drawn onto the image by the user and finally the 
user is prompted to enter the number of osteoclasts (see Note 9) and osteo- 
blasts present in this field. We usually measure at least 10 fields from a repre- 
sentative area of a section, three sections at different levels (at least 100 pm 
apart) per animal and at least six animals per treatment group (see Note 10). 

4. Notes 

1. Mice aged from several days up to several months of age can be used for this 
assay. Neonates require smaller amounts of injection material (useful when using 
an expensive drug) and have the advantage that they are easier to handle. 

2. The injection schedule needs to be optimized for each substance tested in this 
assay. One of the most important variables influencing this is the biological half- 
life of the substance tested. For example, when testing the effects of mevastatin 
we used a regimen of two injections (5 mg/kg) per day for 5 d and euthanize the 
animals 1 d or 7 d after the last injection. 

3. It is essential that all injection solutions and syringes are sterile. Otherwise, the 
effects of a test drug could be masked easily by a localized immune response to 
the injection, which will invariably produce some localized bone loss in the cal- 

4. Embedding in methyl methacrylate (MMA) plastic is not an alternative, as the 
TRAP stain does not work well on material embedded in this plastic. An alterna- 
tive could be the embedding of decalcified calvariae in wax. However, the au- 
thors do not know of any stains for this material that will allow easy distinction 
between bone and the other tissues using simple color thresholding; consequently, 
the semi-automated analysis of these sections will be much more difficult. 

5. We have found that the manufacturer's protocol, which uses only the histocryl 
accelerator, often leads to brittle blocks that are difficult to cut. Our variation, 
which uses an accelerator mix, produces blocks that are easier to cut. 

6. It is essential to perform the embedding at low temperature. Especially the poly- 
merization step. This step is best performed in a crushed ice slush that will opti- 
mally cool the polymerizing block. 

24/van 't Hof/343-352/F1 350 1 2/26/03, 10:48 AM 

I ■ I I I 


Calvarial Injection Assay 351 

7. Celestine Blue can be used as an alternative to hematoxylin if nuclei are not 
stained particularly well. Prepare the Celestine Blue as follows: 2.5 g of Celestine 
Blue B, 25 g of ferric ammonium sulfate, 70 mL of glycerin, and 500 mL of 
dH 2 0. Dissolve the ferric ammonium sulfate in cold distilled water and stir well. 
Add Celestine Blue to this solution, then boil the mixture for a few minutes. 
After cooling, filter the stain and add the glycerin. Use the same staining time for 
Celestine Blue as for hematoxylin, that is, 20 min. 

8. In many programs the calvarial width is determined by having the user draw lines 
across the mineralized bone at multiple sites. This method is fairly time consum- 
ing and not very reproducible owing to operator variability and bias. We use a 
mathematical method, whereby the calvarial bone is modeled as a rectangle and 
the width is calculated from the perimeter and the surface area of the bone ac- 
cording to the following formula: 

Width = Perimeter - VPerimeter 2 - (16 * Area)/4 

To make this method work properly, all holes within the bone binary image 
should be closed (using an Image Holefill operator) and the outline should be 
smoothed by a Binary Close operator (or by an Image Dilate, followed by an 
Image Erode operator of the same size). The aforementioned operators are avail- 
able in all image analysis packages that are currently on the market. 

9. It can be difficult to get an accurate number of osteoclasts, as these cells are often 
present in clusters and not well separated visually. Furthermore, as osteoclasts 
are such large, irregularly shaped cells, what appear to be several osteoclasts 
close together in a section may actually be parts of the same osteoclast. For this 
reason it is good practice to analyze several histological sections, separated by at 
least 100 [im (see Subheading 3.5.). 

10. To avoid possible artefacts introduced by the dissection procedure, do not take 
histomorphometric measurements at the calvarial ends. 


1. Boyce, B. F., Aufdemorte, T. B., Garrett, I. R., Yates, A. J., and Mundy, G. R. 
(1989) Effects of interleukin- 1 on bone turnover in normal mice. Endocrinology 
125, 1142-1150. 

2. van't Hof, R. J., Armour, K. J., Smith, L. M., et al. (2000) Requirement of the 
inducible nitric oxide synthase pathway for IL-1- induced osteoclastic bone 
resorption. Proc. Natl. Acad. Sci. USA 97, 7993-7998. 

3. Mundy, G., Garrett, R., Harris, S., et al. (1999) Stimulation of bone formation in 
vitro and in rodents by statins. Science 286, 1946-1949. 

24/van 't Hof/343-352/F1 351 1 2/26/03, 10:48 AM 

I ■ I I I 


24/van 't Hof/343-352/F1 


2/26/03, 10:48 AM 

I ■ I I I 



Inflammation-Induced Osteoporosis 

The IMO Model 

Kenneth J. Armour and Katharine E. Armour 

1. Introduction 

Generalized osteoporosis and an increased risk of fracture are commonly 
observed in chronic inflammatory diseases such as rheumatoid arthritis, 
ankylosing spondylitis, and inflammatory bowel disease (1-4). Current evi- 
dence suggests that the osteoporosis developed during chronic inflammation 
may result from the inhibition of bone formation, and is associated with sys- 
temic overproduction of proinflammatory mediators, such as cytokines, nitric 
oxide (NO), and prostaglandins (5-8). 

The first animal model of generalized osteoporosis resulting from inflam- 
mation that closely resembled the chronic inflammatory bone loss seen in 
human patients was developed by Minne et al. (9) in the rat. The inflammation- 
mediated osteoporosis (IMO) model utilizes the subcutaneous injection of non- 
specific irritants, such as talc or cotton wool, typically on the back of the rat at 
sites distant from the skeleton, to stimulate an acute phase response. Granulo- 
matous reactions are noted at the injection sites along with the accumulation of 
chronic inflammatory cells. In the skeleton, at sites distant from the inflamma- 
tory lesions, loss of trabecular bone volume and significant decreases in osteo- 
blast numbers are observed (Fig. 1). Subsequent studies have extended these 
observations to show that decreases in osteoblast numbers and bone formation 
are major features of the IMO model (10-12). In growing rats, decreases in 
osteoprogenitor number and bone elongation are also observed (12). Interest- 
ingly, osteoclast numbers and osteoclastic resorption are generally unchanged 
(or transiently decreased) in this model (9-13), which also accords with the 
observations in humans with chronic nonosseous inflammation (1-4). 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


25/Armour/353-360/F1 353 1 2/26/03, 10:49 AM 

I ■ I I I 


Armour and Armour 



Fig. 1 . Photomicrographs of distal femurs from rats 2 1 d after treatment with saline 
(A) or talc (B). Trabecular thinning and lower numbers of trabeculae are evident in 
(B). (Reduced from original magnification, x25.) 

Elevated tumour necrosis factor-a (TNF-a) levels have been shown to play 
a causal role in the bone loss seen in the IMO model and the effects of IMO can 
be neutralized with anti-TNF-a antibodies (14). Furthermore, TNF-a is known 
to stimulate NO production. Recent work has shown that increased osteoblast 
apoptosis, associated with the production of high levels of NO synthesized by 
the inducible nitric oxide synthase (iNOS) pathway (15,16), contributes to the 
pathogenesis of IMO. 

This chapter describes a general protocol for the induction of IMO in rodents 
(rats and mice) using injections of talc and suggests suitable end points for the 
experimental investigations. 

2. Materials 

1. Sterile isotonic (0.9%) saline (Sigma, Poole, UK) for use as an injection vehicle. 
Phosphate-buffered saline or a balanced salt solution (such as Hanks') can be 
used as alternatives. 

2. Talc (hydrous magnesium silicate; Sigma), sterilized by heating at 160°C for at 
least one hour (17). After sterilization, aseptically prepare an 800 mg/mL suspen- 
sion in sterile saline. Store at room temperature and use shortly after preparation. 
The talc suspension can be stored at 4°C for short periods but needs to be warmed 
to room temperature and mixed thoroughly before use. 

3. Calcein (Sigma), prepared as a 4 mg /mL solution in sterile saline and adjusted to 
around neutral pH. Filter sterilize using a 0.2-jxm filter (Acrodisc; Pall, Ports- 
mouth, UK), then aliquot. Protect aliquots from light during storage, preferably 
at -20°C if stored for extended periods. 

4. Vetalar (ketamine hydrochloride: 100 mg/mL; Pharmacia & Upjohn Ltd., 
Crawley, UK) and Rompun (xylazine: 2% solution; Bayer pic, Bury St. Edmonds, 
UK) for the induction of anesthesia. 



2/26/03, 10:49 AM 

I ■ I I I 


Inflammation-Induced Osteoporosis 355 

5. 25-Gauge needles (BD UK Ltd., Cowley, UK). 

3. Methods 

3. 1. Animals 

1 . Ideally, at least 10 mice or rats of the same sex should be used per group to ensure 
that any differences or changes measured reach statistical significance (see Notes 
1 and 2). 

3.2. Inflammation-Mediated Osteoporosis 

The following protocol describes a 21-d longitudinal study. 

1 . For urine collection, place animals in individual metabolic cages prior to the start 
of experiment and collect the urine samples into sterile containers over a 16-24-h 
period. Add a small volume of 5 M sodium hydroxide to each sample to prevent 
bacterial growth and freeze the samples immediately. Snap freezing in liquid 
nitrogen is preferable. 

2. Weigh all the animals (to at least one decimal place for mice) (see Note 3). 

3. Anesthetize animals with Vetalar (100 mg/kg of body weight) and Rompun (20 
mg/kg body weight) by intraperitoneal (i.p.) injection (into the lower left quad- 
rant of the abdomen). This will induce medium-depth anesthesia of sufficient 
duration to perform the subsequent two experimental steps (see Notes 4 and 5). 

4. Using a 25-gauge needle, inject the talc suspension subcutaneously at four or 
more sites on the upper back and sides of the animal, to give a total talc dose of 
16 mg/g body weight. Administer an equivalent volume of sterile saline to the 
controls. Avoid injecting directly over skeletal sites. 

5. Scan the animals using bone densitometry. Dual-energy X-ray Absorptiometry 
(DXA) is useful for detecting bone mass changes over the whole skeleton, but 
peripheral quantitative computed tomography (pQCT) is preferable for assessing 
discrete changes in trabecular and cortical bone components (see the chapter by 
Gasser, this volume, for a full discussion of pQCT) (see Notes 6 and 7). 

6. Administer calcein (40 mg/kg body weight) by i.p. injection at two or more time 
points during the IMO procedure to label bone fluorescently. In our laboratory 
we have obtained useful information from calcein injections on d 10 and d 17. 
The timing of the calcein injections is important and should be standardized to 
the same time of day on each of the injection days. Calcein is preferred over 
tetraxyclin in rodents. 

7. Repeat experimental step 1 on the penultimate day of the 21-d experimental 

8. On the final day of the experiment repeat experimental steps 2, 3, and 5. 

9. Euthanize the animals. 

10. Remove the right and left hind leg bones (femorae and tibiae) and fix in prepara- 
tion for bone histomorphometry (see Table 1 and Note 8) (see the chapter by 
Vedi and Compston, this volume, for details). 

11. Remove and weigh the spleens (see Note 9). 

25/Armour/353-360/F1 355 1 2/26/03, 10:49 AM 

I ■ I I I 

356 Armour and Armour 

Table 1 

Overview of the Pathophysiological Changes in Bone 

Associated with Inflammation-Mediated Osteoporosis 

Bone parameter 



Trabecular bone 






Osteoblast numbers 



Osteoblast apoptosis 
Bone formation 



Bone elongation/growth 
Osteoclast numbers 


•**> or U (transient) 


Bone resorption 

<t* or |J. (transient) 



4. Notes 

1. Rats weighing 200 g or more and mice older than 14 wk of age are preferable. 
The potentially confounding effects of skeletal growth need to be accounted for 
when assessing the effects of IMO in younger animals. 

2. Age matching is important, particularly when studying bone changes in 
transgenic mice, and the use of littermates is recommended. 

3. It has been noted that talc-injected rats fed ad libitum consume less food and 
exhibit less weight gain (or even some weight loss) compared with saline-in- 
jected rats. For this reason it may be desirable to match and pair-feed controls 
with talc-injected rats to minimize any effect of lower caloric intake on bone 
(11,12,14,17). Rats should be housed in individual cages to allow for this feeding 
regimen. We have not observed the same tendency to lower food consumption in 
mice following talc injection. 

4. For anesthesia of mice, it is practical to prepare 1:10 dilutions of Vetalar and 
Rompun in sterile saline, using the same container for both dilutions to minimize 
the injection volume required. The administration of 0.1 mL/10 g of body weight 
provides the required dose (100 mg of Vetalar/kg body weight and 20 mg of 
Rompun/kg of body weight). In rats, 0.1 mL of each of the undiluted Vetalar and 
Rompun stocks should be used per 100 g of body weight to provide the correct 
dose for anesthesia. The solutions can be stored at 4°C. 

5. It may be desirable to take small blood samples during the IMO experiment, for 
example, to measure circulating cytokine or electrolyte levels. In our hands, the 
use of a Vetalar-Rompun mix provides reliable and predictable anesthesia and is 
not associated with significant animal mortality. In practice, however, Vetalar- 
Rompun-induced anesthesia tends to decrease the amount of blood obtainable 
from sampling, and inhalation anesthesia using halothane is preferable in this 
situation. Induction of anesthesia using 3-4% halothane followed by the mainte- 

25/Armour/353-360/F1 356 1 2/26/03, 10:49 AM 

I ■ I I I 


Inflammation-Induced Osteoporosis 357 

nance of anesthesia using 1-2% halothane in the presence of an adequate flow of 
oxygen is effective for blood sampling. A 2-wk interval between blood sam- 
plings is generally recommended (18), and samples taken at the start and end of 
the experiment are feasible when using a 21-d protocol. Blood withdrawal from 
the lateral vein of the tail is a minimally invasive method and recommended for 
small samples. Terminal blood samples at the end of the experiment can be also 
obtained by cardiac puncture, and inhalation anesthesia is again preferable for 
this. The use of inhalation anesthesia does not preclude the use of an injectable 
anesthetic and the latter may still be appropriate when subsequently undertaking 
densitometric measurements. Note that only fully trained personnel, authorized 
by their relevant institutional and national authorities, should undertake proce- 
dures on animals. 

6. Longitudinal studies are clearly preferable to cross-sectional studies because of 
the additional experimental information they provide and the lower number of 
animals they require. However, if an end point only study is to be conducted, it is 
essential that an additional group of animals, killed at the start of the experiment, 
is included to provide baseline readings. 

7. A 21-d period is ideal for demonstrating the osteoporosis that results from gener- 
alized inflammation and for investigating the bone-sparing effects of potential 
therapies. Although the magnitude of trabecular bone loss may be greater in 
longer experiments, studies of shorter duration have also been used successfully, 
with osteoblast insufficiency and some trabecular bone loss evident within 7 d of 
talc administration (12,19). 

8. It is good practice to obtain as much information as possible from each in vivo 
experiment and biomechanical and micro-CT analyses of bone (discussed in the 
chapters by Aspden and by Gasser, respectively, this volume) should be seriously 
considered as additional options when assessing the effects of experimentally 
induced bone loss. For biomechanical studies, freshly dissected bones should 
tested as soon as possible after dissection; alternatively, the bones should be 
wrapped in gauze or tissue paper soaked with sterile saline and frozen at -20°C 
in sterile containers until required. 

9. As listed in Table 1, IMO is associated with marked pathological changes in 
bone that include decreased trabecular bone mass, osteoblast insufficiency, and 
decreased bone formation. In addition to pathological changes in bone, it is im- 
portant to have bone-independent measures of the effectiveness of talc-induced 
inflammation. In this regard, the following indices may be applicable: 

a. Splenomegaly is a clear and useful indicator of an inflammatory response 
(10) and weighing a dissected spleen is simple to perform. A 400% increase 
in spleen weight in mice compared with controls is not uncommon (16). 

b. Microscopic examination of the spleen may also be useful for describing the 
cellular changes associated with the acute phase response (20) . The spleens should 
be fixed or frozen in liquid nitrogen prior to immunohistological analysis. 

c. Urinary nitrate excretion is elevated during IMO as a result of increased NO 
production and the measurement of urinary NO.r using the Griess reaction 
(21) is another useful and straightforward indicator of inflammation. 

25/Armour/353-360/F1 357 I 2/26/03, 10:49 AM 

I ■ I I I 


Armour and Armour 

Table 2 

Overview of the Pathophysiological Changes Associated with 

Inflammation-Mediated Osteoporosis 




Serum components/trace elements 






















Serum proteins 




Alkaline phosphatase (liver isoform) 



C-reactive protein 



a r , a 2 -, and (3-globulin 






Hormones/growth factors 

Adrenocorticotropic hormone 






Parathyroid hormone 






1,25-Dihydroxy vitamin D^ 

|| (transient) 








Urinary excretion 




Nitrate/nitrite (NO*) 



White blood cell counts 

Mononuclear cells 



Polymorphonuclear cells 



Table 2 lists numerous other humoral or cellular changes that have been 
reported in response to IMO, and many of these can be used as bone-indepen- 
dent indicators of the inflammatory response. 


1. Deodhar, A. A. and Woolf, A. D. (1996) Bone mass measurement and bone 
metabolism in rheumatoid arthritis: a review. Br. J. Rheumatol. 35, 309-322. 

2. Andreassen, H., Rungby, J., Dahlerup, J. F., and Mosekilde, L. (1997) Inflamma- 
tory bowel disease and osteoporosis. Scand. J. Gastroenterol. 32, 1247-1255. 



2/26/03, 10:49 AM 

I ■ I I I 


Inflammation-Induced Osteoporosis 359 

3. Will, R., Palmer, R., Bhalla, A. K., Ring, F., and Calin, A. (1989) Osteoporosis in 
early ankylosing spondylitis: a primary pathological event? Lancet 2, 1483-1485. 

4. Ralston, S. H., Urquhart, G. D. K., Brzeski, M., and Sturrock, R.D. (1990) Preva- 
lence of vertebral compression fractures due to osteoporosis in ankylosing 
spondylitis. Br. Med. J. 300, 563-566. 

5. Oudkerk, P. M., Bouma, G., Visser, J. J., et al. (1995) Serum nitrate levels in 
ulcerative colitis and Crohn's disease. Scand. J. Gastroenterol. 30, 784-788. 

6. Sakurai, H., Koshaka, H., Lui, M-F., et al. (1995) Nitric oxide production and 
inducible nitric oxide synthase expression in inflammatory arthritides. /. Clin. 
Invest. 96, 2357-2363. 

7. Dijkstra, G., Moshage, H., van Dullemen, H. M., et al. (1998) Expression of nitric 
oxide synthases and formation of nitrotyrosine and reactive oxygen species in 
inflammatory bowel disease. /. Pathol. 186, 416-421. 

8. Ralston, S.H. (1997) Nitric oxide and bone: what a gas! Br. J. Rheumatol. 36, 831-838. 

9. Minne, H. W., Pfeilschifter, J., Scharla, S., et al. (1984) Inflammation-mediated 
osteopenia in the rat: a new animal model for pathological loss of bone mass. 
Endocrinology 115, 50-54. 

10. Pfeilschifter, J., Wuster, C, Vogel, M., Enderes, B., Ziegler, R. and Minne, H. W. 
(1987) Inflammation-mediated osteopenia (IMO) during acute inflammation in rats 
is due to a transient inhibition of bone formation. Calcif. Tissue Int. 41, 321-325. 

11. Vukicevic, S., Marusic, A., Stavljenic, A., Cicak, N., Vogel, M., and Krempien, 
B. (1988) Talc granulomatosis in the rat: the relationship between osteoblast 
insufficiency and adjacent bone marrow hyperplasia. Exp. Hematol. 16, 735-740. 

12. Krempien, B., Vukicevic, S., Vogel, M., Stavljenic, A., and Buchele, R. (1988) 
Cellular basis of inflammation-induced osteopenia in growing rats. /, Bone Miner. 
Res. 3, 573-582. 

13. Lempert, U. G., Minne, H. W., Fleisch, H., Muhlbauer, R. C, Scharla, S. H., and 
Ziegler, R. (1991) Inflammation-mediated osteopenia (IMO): no change in bone 
resorption during its development. Calcif. Tissue Int. 48, 291-292. 

14. Vukicevic, S., Marusic, A., Stavljenic, A., Cesnjaj, M., and Ivankovic, D. (1994) 
The role of tumor necrosis factor-alpha in the generation of acute phase response 
and bone loss in rats and talc granulomatosis. Lab. Invest. 70, 386-391. 

15. Armour, K. E., van 't Hof, R. J., Grabowski, P. S., Reid, D. M. and Ralston, S. H. 
(1999) Evidence for a pathogenic role of nitric oxide in inflammation-induced 
osteoporosis. /. Bone Miner. Res. 14, 2137-2142. 

16. Armour, K. J., Armour, K. E., van 't Hof, R. J., et al. (2001) Activation of the 
inducible nitric oxide synthase pathway contributes to inflammation-induced 
osteoporosis by suppressing bone formation and causing osteoblast apoptosis. 
Arthrit. Rheum. 44, 2790-2796. 

17. Marusic, A., Kos, K., Stavljenic, A., and Vukicevic, S. (1991) Acute zinc defi- 
ciency and trabecular bone loss in rats with talc granulomatosis. Biol. Trace. Elem. 
Res. 29, 165-173. 

18. Waynforth, H. B. and Flecknell, P. A., eds. (1998) Experimental and Surgical 
Techniques in the Rat, 2nd edit. Academic Press, London. 

25/Armour/353-360/F1 359 1 2/26/03, 10:49 AM 


I ■ I I I 

360 Armour and Armour 

19. Marusic, A., Kos, K., Stavljenic, A., and Vukicevic, S. (1990) Talc granulomato- 
sis in the rat. Involvement of bone in the acute-phase response. Inflammation 14, 

20. Radic, I., Vucak, I., Milosevic, J., Marusic, S., Vukicevic, S., and Marusic, M. 
(1988) Immunosuppression induced by talc granulomatosis in the rat. Clin. Exp. 
Immunol. 73, 316-321. 

21. Green, L. C, Wagner, D. A., Glogowski, J, Skipper, P. L., Wishnock, J. S., and 
Tannenbaum, S. R. (1982) Analysis of nitrate, nitrite and (15N) nitrate in biologi- 
cal fluids. Analyt. Biochem. 126, 131-138. 

22. Hadjidakis, D., Lempert, U. G., Minne, H. W. and Ziegler, R. (1993) Bone loss in 
experimental diabetes. Comparison with the model of inflammation mediated 
osteopenia. Horm. Metab. Res. 25, 77-81. 

23. Marusic, A., Kos, K., Stavljenic, A., and Vukicevic, S. (1993) Role of 1,25- 
dihydroxyvitamin D3 in the generation of the acute-phase response in rats with 
talc-induced granulomatosis. Experientia 49, 693-698. 

25/Armour/353-360/F1 360 1 2/26/03, 10:49 AM 

I ■ I I I 



Ovariectomy and Estrogen Replacement in Rodents 

Jade W. M. Chow 

1 . Introduction 

Estrogen is known to be one of the major hormonal influences in bone 
remodeling and bone mass. Estrogen deficiency after the menopause is one of 
the leading causes of osteoporosis, and currently estrogen replacement is the 
first line management for postmenopausal osteoporosis. The bone loss associ- 
ated with estrogen deficiency is due to increased bone resorption and a relative 
deficiency in bone formation. Although estrogen is thought to prevent bone 
loss mainly by suppressing bone resorption (1,2), there is also recent evidence 
to suggest that estrogen may exert an anabolic effect in bone in humans (3,4). 
Estrogen receptors (ER) are present in osteoblasts (5), and oestradiol has been 
shown to increase type I collagen and alkaline phosphatase production by 
osteoblasts in vitro (6). Animal models have proved invaluable in the study of 
the role of estrogen in bone metabolism. There is a large body of evidence that 
the cancellous bone of the secondary spongiosa of adult female rats has char- 
acteristics similar to that of humans. The secondary spongiosa undergoes bone 
remodeling and becomes osteopenic with disuse and estrogen deficiency. Ova- 
riectomy induces increased bone resorption that in turn entrains increased bone 
formation. Resorption, formation and bone loss are all suppressed by estrogen, 
calcitonin, and bisphosphonates. Preliminary in vivo studies in the evaluation 
of new compounds directed at the estrogen receptor are generally performed in 
rats. For this reason, this chapter concentrates on ovariectomy and estrogen 
replacement in this species. With the development of transgenic and gene dele- 
tion animals, there is an increased demand for similar experiments in mice. 
While the bones of these animals behave in a similar manner to rats in many 
respects, its marked anabolic response to estrogen replacement is unusual and 
controversial (7,8). 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


26/Chow/361 -366/F1 361 1 2/26/03, 10:49 AM 

I ■ I I I 


362 Chow 

2. Materials 

1. Estrogen: Prepare estradiol (Sigma) by sonication in a vehicle of 5% v/v benzyl 
(Sigma) and 95% v/v corn oil (Sigma). This may be stored for up to 4 wk at room 
temperature away from light. 

2. Estrogen slow release pellets: l\ig and 10u,g, Innovative Research of America, FL. 

3. Dissecting Instruments. Selection of scissors, scalpels, and forceps. 

4. Silk sutures (Ethicon). 

5. Metal clips for wound closure (Brookwick ward, Fife, Scotland). 

6. Mayer's hematoxylin and eosin: For staining vaginal smears (BDH, Poole). 

7. Ethanol series: 25%, 50%, 75%, 90%, 100%. To dehydrate vaginal smear specimens. 

8. Xylene. To dehydrate vaginal smear specimens. 

9. DePeX mounting medium: For mounting vaginal smear specimens (BDH, Poole). 

3. Methods 

3. 1. Ovariectomy 

1. Anesthetize the animal with nitrous oxide and oxygen at 1 L/min and 2 L/min 
respectively and 2% v/v halothane (Sigma). 

2. In rats, the dorsal aspect of each animal is first shaved. In mice, the fur of the 
dorsal surface is wetted by ethanol, and then gently parted to create a center line 
for the skin incision. 

3. Make a midline dorsal incision (approx 2 cm in the rat) using a pair of sharp 
scissors. The subcutaneous connective tissue on either side can then be freed 
from underlying muscle by blunt dissection using a sterile pair of blunt artery 
forceps. A smaller incision (<1 cm in the rat) in the muscle layer on either side of 
this is then made to allow entry into the peritoneal cavity. 

4. The ovaries can be identified in a fat pad adjacent to the perinephric region and 
should be excised as a whole. In mice, ovaries are sometimes visible even with- 
out the muscle incision because the muscle layer is quite thin. Removal of the 
fimbrial end of the fallopian tube may be done to ensure completeness of the 

5. Suture the muscle incision with silk, and close the skin incision with a metal clip. 

6. Keep the animals under observation until they have recovered from the anes- 
thetic and regained consciousness (see Note 1). 

7. The same procedures are followed for sham operations except that the ovaries are 
identified but not removed. 

3.2. Estrogen Injections 

Prepare the estrogen as a stock solution so that the animal does not receive 
more than 1 mL/kg body weight to achieve the required amount of compound. 
The injections may be given daily, once, twice, or three times per week depending 
on the experimental design (see Note 2). For experiments that involve daily 
injections over prolonged periods, the injection sites should be rotated. 

26/Chow/361 -366/F1 362 I 2/26/03, 10:49 AM 

I ■ I I I 


Estrogen Replacement in Rodents 363 

3.3. Estrogen Implants 

Estrogen replacement and vehicle may also be given subcutaneously as slow 
release pellets. The pharmokinetics of the various forms of exogenous replace- 
ments are likely to be different, and this may need to be considered in the experi- 
mental design. They have not been studied in detail, however, (see Note 2). 

3.4. Ovarian Transplantation 

These may be done as renal capsule ovarian transplants or subcutaneous 
transplants (9,10) (see Note 3). 

1. For renal ovarian transplants, adult (13 wk old) donor rats should be anesthetized 
and the ovaries exposed as described in Subheading 3.1. 

2. Reflect the ovary onto a piece of sterile gauze and under a dissecting microscope, 
slit the ovarian bursa, and cut the hilum to release the ovary. 

3. The recipient ovariectomized host animals should be simultaneously anesthetized. 

4. For a renal capsule transplant, the left kidney is exposed through a dorsal skin 
and muscle incision and a small slit made in the renal capsule covering the kidney. 

5. The donor ovary is cut into three segments to facilitate insertion and these are 
gently place these under the renal capsule and sutured in place using fine silk. 

6. For subcutaneous transplantation, the ovarian segments are inserted into a subcu- 
taneous pocket made using a skin incision followed by blunt dissection in the 
mid-scapular region. 

7. When the transplant is complete, the skin incision should be closed with a metal clip. 

3.5. Age of Animals and Duration of Follow-Up 

This varies according to the study design and question being asked. How- 
ever studies of ovariectomy induced bone loss in rats are typically carried out 
in 12-16-wk-old adult females, with follow-up for 6 wk, by which time there 
will have been significant bone loss which can be detected by 
histomorphometry (see the chapter by Vedi and Compston, this volume) or 
bone densitometry (see the chapter by Gasser, this volume). Experiments in 
mice are usually performed in 8-12-wk-old females with follow up for 3 wk. 

3.6. Checking Efficacy of Ovarian Transplants Using Vaginal Smears 

This is a simple and cheap method for testing the efficacy of ovarian transplants, 
and can be used in addition to monitoring estrogen and progesterone levels. 

1. Moisten a small sterile cotton wool bud with normal saline solution, insert it into 
the vagina, and gently scrape the vaginal wall to obtain some cells. 

2. Smear the cells onto a glass slide, air-dry, and fix with 70% ethanol. 

3. Stain the cells for 5 min in Mayer's hematoxylin. 

4. Wash in running water, and counterstain for 2 min in eosin. Rinse in running water. 

26/Chow/361 -366/F1 363 1 2/26/03, 10:49 AM 

I ■ I I I 



Table 1 

The Four stages of the Rat Estrous Cycle, Their Duration, and Associated 

Changes in Vaginal Cytology 

Approximate C y tol °gy 

Stage of cycle duration (h) Epithelial cells Cornified epithelial cells Leukocytes 

+ ± 

+ ++ 















- None; ± occasional; + few; ++ many; +++ abundant.The laboratory rat has a regular estrous cycle 
of approx 4-5 d duration. The cycle is divided into four distinct stages based on microscopic examina- 
tion of cell morphology of the vaginal smear. 

5. Dehydrate the slides through graded ethanols (in the order of 25%, 50%, 75%, 
90%, and 100%), for 5 min each. 

6. Dip the slides in xylene and mount in DePeX. 

7. Examine the slides microscopically for the presence of leucocytes, epithelial 
cells, and cornified epithelium. The four stages of the estrous cycle can be recog- 
nized by results of microscopic examination of the vaginal smear (Table 1). 

4. Notes 

1 . The animals should be mobile and feed freely within 30 min. 

2. The pharmacokinetics of a single subcutaneous injection of 40 p.g/kg estradiol is 
such that peak serum levels are achieved after 2 h and the half-life is 8 h. Serum 
estrogen returns to near basal levels by 24 h after the injection (11). The usual 
replacement dose of estradiol for rats is 5 (xg/kg/d. Although this restores serum 
estradiol levels to those seen in intact cycling animals, and suppresses bone loss, 
the uterus weight tends to be lower than in normal animals, indicating that injec- 
tions do not replicate the normal physiological state. 

3. Both methods of ovarian transplantation are fully efficacious in restoring estro- 
gen levels to that of sham-operated animals, but renal capsule ovarian transplan- 
tation is slightly more successful in restoring ovarian function if progesterone 
levels and number of normal oestrous cycles are also considered. This is due 
possibly to increased vascular supply associated with the renal capsule. For most 
bone studies, both methods are equivalent, as progesterone does not appear to 
play a major role in bone metabolism. 


1. Vedi, S. and Compston, J. E. (1996) The effects of long-term hormone replace- 
ment therapy on bone remodeling in postmenopausal women. Bone 19, 535-539. 

2. Steiniche, T., Hasling, C, Charles, P., Eriksen, E. F., Mosekilde, L., and Melsen, 
F. (1989) A randomized study on the effects of estrogen/gestagen or high dose 

26/Chow/361 -366/F1 


2/26/03, 10:49 AM 

I ■ I I I 


Estrogen Replacement in Rodents 365 

oral calcium on trabecular bone remodeling in postmenopausal osteoporosis. Bone 

3. Khastgir, G., Studd, J., Holland, N., Alaghband-Zadeh, J., Fox, S., and Chow, J. 
(2001) Anabolic effect of estrogen replacement on bone in postmenopausal 
women with osteoporosis: histomorphometric evidence in a longitudinal study. /. 
Clin. Endocrinol. Metab. 86, 289-295. 

4. Vedi, S., Purdie, D. W., Ballard, P., Board, S., Cooper, A. C, and Compston, J. E. 
(1999) Bone remodeling and structure in postmenopausal women treated with 
long-term, high-dose estrogen therapy. Osteoporosis Int. 10, 52-58. 

5. Komm, B. S., Terpening, C. M., Benz, D. J., et al. (1988) Estrogenic binding, 
receptor mRNA, and biologic response in osteoblast-like osteosarcoma cells. Sci- 
ence 241, 81-84. 

6. Ernst, M., Heath, J. K., and Rodan, G. A. (1989) Estradiol effects on proliferation, 
messenger ribonucleic acid for collagen and insulin-like growth factor-I, and par- 
athyroid hormone-stimulated adenylate cyclase activity in osteoblastic cells from 
calvariae and long bones. Endocrinolog 125, 825-833. 

7. Samuels, A., Perry, M. J., and Tobias, J. H. (1999) High-dose estrogen induces 
medullary bone formation in female mice. /. Bone Miner. Res. 14, 178-186. 

8. Turner, R. T. (1999) Mice, estrogen and postmenopausal osteoporosis. /. Bone 
Miner. Res. 14, 187-191. 

9. Parkes, A. S. (1956) Survival time of ovarian homografts in two strains of rats. /. 
Endocrinol. 13, 201-210. 

10. Felicio, L. S., Nelson, J. F., Gosden, R. G., and Finch, C. E. (1983) Restoration of 
ovulatory cycles by young ovarian grafts in aging mice: potentiation by long-term 
ovariectomy decreases with age. Proc. Natl. Acad. Sci. USA 80, 6076-6080. 

11. Jagger, C. J., Chow, J. W. M., and Chambers, T. J. (1996) Estrogen suppresses 
activation but enhances formation phase of osteogenic response to mechanical 
stimulation in rat bone. /. Clin. Invest. 98, 2351-2357. 

26/Chow/361 -366/F1 365 1 2/26/03, 10:49 AM 

I ■ I I I 


26/Chow/361 -366/F1 


2/26/03, 10:49 AM 

I ■ I I I 



Mechanical Testing of Bone Ex Vivo 

Richard M. Aspden 

1. Introduction 

The primary function of bone is to form the skeleton, which provides sup- 
port for the body and protection for vital organs. These are primarily mechani- 
cal functions. To fulfil these, the bone matrix has to have the right combination 
of stiffness and strength to enable it to withstand the forces imposed upon it. 
These forces may be repetitive and moderate, such as those generated during 
walking, or high and transient, such as inflicted by a blow on the head. The 
structure and composition of bone can adapt over time to try match the 
mechanical properties of the bone to the prevailing demands being placed on it. 
How to measure some of these mechanical properties is the aim of this chapter. 

There are two types of bone that are considered in this chapter: cortical, or 
dense, bone and cancellous, or trabecular, bone. Cortical bone is a solid, com- 
pact material that forms the diaphyses of long bones, and a shell around the 
metaphyses and the vertebrae. Cancellous bone has an open, porous structure 
comprising rods or plates. Hence it is less dense than cortical bone and also 
less stiff and strong. It makes up the center of the vertebrae and the metaphy- 
ses. However, the two materials in combination form structures that are strong 
and light and their properties in situ may differ considerably from those mea- 
sured from tests on isolated samples (1,2). In cancellous bone the distinction 
between material and structure is not always easy to define, and treating it as a 
cellular material has met with a considerable degree of success (3,4). 

Although conceptually relatively simple, the mechanical properties of these 
tissues are not easy to measure. If only a comparison is required between 
equivalent sites in different patient groups then careful use of the same tech- 
nique on all samples will provide consistent relative values. If, however, abso- 
lute values are required then great care must be taken as bone is not isotropic or 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


27/Aspden/367-380/F1 369 1 2/26/03, 10:49 AM 

I ■ I I I 


370 Aspden 

homogeneous and its properties depend on the rate at which it is deformed. In 
addition, sample preparation can be difficult. Unlike engineering materials, 
which can be cut and machined to predetermined sizes for testing and for which 
there are officially recognized standards, bone samples are generally limited in 
size and shape by the site from which they are taken. The size of the sample 
being tested may affect the measurements being made. 

This chapter describes methods of preparation and testing that can be applied 
to cortical and cancellous bone. Differences in approach that are required 
because of the different natures of the materials are noted. The descriptions are 
restricted to methods that can reasonably easily be applied by anyone having 
access to a materials testing machine. Researchers requiring more sophisti- 
cated techniques or more details of variations are referred to the recent book by 
Stephen Cowin (5) and to the original papers. Finally, a method is presented 
for testing whole bones, such as may be required to measure bone properties in 
genetically modified mice. These are so small that preparing and testing iso- 
lated samples of cortical or cancellous bone becomes very difficult and tests 
are generally performed on intact bones. 

2. Materials 

1. Phosphate-buffered saline (PBS) (Gibco): Used to keep specimens moist during 
preparation and testing (see Note 1). 

2. Hacksaw: For cutting large bone samples to manageable sized pieces prior to 
precision cutting. 

3. Mineralogical saw fitted with either a diamond or an aluminum oxide cutoff 
wheel. To precision cut bone samples (see Note 2). 

4. Lathe or Milling machine: To mill bone samples to desired shape. 

5. Coring bits, 5-mm and 9-mm internal diameter and break-off tools: Custom made 
by Bolton Surgical Ltd. (Sheffield, UK). 

6. Materials testing machine: For mechanical testing of samples. 

7. Software for calculation of biomechanical variables: Origin (OriginLab Corp., 
Northampton, MA, USA). 

3. Methods 

3. 1. General Considerations for the Preparation of Bone Samples 

All human tissue must be handled with due regard to health and safety be- 
cause of the risk of infection. In addition to gloves and laboratory coat, a mask 
and eye shield are required because there is a risk of aerosols being generated 
during cutting, drilling, or machining. Samples must be kept moist at all times 
during the preparation procedure. Bones must be dissected free of soft tissues, 
taking care not to create notches in the bone that is going to be tested, as this 
will weaken it. General methods for the preparation of samples for testing are 
described in the following sections. 

27/Aspden/367-380/F1 370 1 2/26/03, 10:49 AM 

I ■ I I I 

Mechanical Testing of Bone Ex Vivo 


Fig. 1. Tensile test specimens are traditionally produced with a waist to ensure that 
failure consistently occurs in the center and not near the grips. 


3.2. Preparing Prisms of Cortical Bone 

1. Cut the bone into manageable sized pieces using a hacksaw or a junior hacksaw. 

2. Precision cut the bone samples to the desired size using a mineralogical saw ro- 
tating at 600-800 rpm. For compression and bending tests the samples should be 
cut to give rectangular parallelepiped shaped specimens. For tensile testing, the 
specimens should be milled to give a waisted configuration (Fig. 1). If the 
samples are large enough, they should have cylindrical symmetry, but for thin 
specimens the waisting can be done in two dimensions. This is to ensure that 
fracture occurs consistently in the central part of the specimen away from the 
gripped ends. Water should be run over the sample at all times during sawing and 

3.3. Preparing Cancellous Bone Cores 

1 . Obtain the bone core by drilling through the bone sample with a coring bit (Fig. 2). 
To remove the core it either has to be drilled right through a piece of bone or broken 
off the underlying bone. Providing the core is not too deep, this can be done using 
the break-off tool shown in Fig. 2. This is made to the same specification as the 
coring bit except it does not require teeth and half of its circumference has been 
removed. This is inserted in place of the coring bit and, by levering on the end of it, 
the base of the core may be broken from the underlying tissue (see Note 3). 

2. Trim the ends of the cores plane and parallel using a mineralogical saw. If the cores 
have been obtained from the articular surface of joints, they will have cartilage and 
subchondral bone at one end that needs to be removed using the saw. 



2/26/03, 10:49 AM 

I ■ I I I 




Fig. 2. Coring bits and matching "break-off devices. Unless cores can be drilled 
right through a piece of bone they will need to be removed from the underlying bone to 
which they are attached. Inserting the break-off device after drilling the core and 
levering it sideways generally successfully snaps the core at its base, after which it 
may be removed using forceps. 

3.4. General Considerations for Mechanical Testing 

1. Before testing it is important to decide what parameters are going to be mea- 
sured. Most testing machines apply a deformation to a sample, which normally 
increases linearly with time. This deformation is commonly referred to as dis- 
placement. The operator can choose the rate at which the displacement occurs 
and the properties measured will depend on this rate as bone is viscoelastic. 
Choosing when to stop the test is also important depending on what is to be 
done next with the specimen. In tension or bending the machine can be set to 
stop automatically on fracture, as the load drops rapidly at this point. In com- 
pression, failure is not so easy to determine and various methods can be used. 
The applied load and displacement are recorded throughout the test. The load- 
displacement curve represents the extrinsic properties of the specimen under 
test and is therefore used to measure the properties of whole bones (Fig. 3A). 
These properties are important for understanding why bones fracture. For pre- 
pared samples of known dimensions what is commonly required are the intrin- 
sic properties of the bone material itself. To obtain these, the load is divided by 
the cross-sectional area of the specimen to produce the stress, and the displace- 
ment is divided by the original length to give the strain. The stress-strain curve 
is then very similar to the load-displacement curve but refers now to the mate- 
rial not the structure (Fig. 3B). 



2/26/03, 10:49 AM 

I ■ I I I 

Mechanical Testing of Bone Ex Vivo 








f 1 




E=c/e | 

/ e 



/ elastic 



Fig. 3. A load-displacement curve (a) measures the extrinsic properties of a speci- 
men, for example, an individual bone. The main parameters are stiffness, work to 
failure (shaded) and ultimate load and displacement. The stress-strain curve (b) is similar 
but, because it is normalized for the sample dimensions, measures the intrinsic properties 
of the material being tested. These are the elastic or Young's modulus (£), the yield stress 
and strain, the ultimate stress and strain, and the energies to yield and failure. 

3.5. Repeating Tests on the Same Specimens 

1 . It has been found that if a test is repeated several times on the same specimen 
within a short period then the resulting load-displacement plots do not immedi- 
ately coincide but appear to converge toward a stable curve. Because of this, 
many investigators apply up to five or six loading cycles to a specimen, often at 
a reduced load, before performing the actual test, a process called precondition- 



2/26/03, 10:49 AM 

I ■ I I I 


374 Aspden 

ing. This is a point of some contention because it is not understood what happens 
during this process. It can be argued that what is finally being measured is not the 
natural property of the specimen but one that has somehow been modified by a 
series of cyclically applied loads. Does repeatability necessarily imply accuracy? 
We always apply a single test to a specimen. 

3.6. Types of Mechanical Testing 

There are four main types of test: tension, compression, bending, and torsion. 

1. Tension and compression testing: These are most commonly done to measure the 
intrinsic properties of bone matrix. These tests use machined samples of known 
dimensions. Tension is generally used for cortical bone and compression is more 
common for cancellous bone because of the difficulty obtaining and mounting 
suitably sized specimens for tensile testing. Bending and torsion can be applied 
in this way but the interpretation of the results is more difficult. They are prob- 
ably more commonly applied to whole bones. In this section, descriptions are 
given of tension and compression tests for machined specimens. 

2. Bending: This is most commonly applied to whole long bones from rodents 
because of the difficulties of machining tensile or compressive specimens from 
such small bones. It is best for measuring extrinsic properties of the whole bone 
and therefore is very useful for studies comparing the effects of different drug 
therapies or genetic modifications. It is not easy to estimate intrinsic properties 
from this test for a variety of reasons, mainly related to the size and shape of the 
bone. More details are given by Turner and Burr (8). There are two bending 
configurations, called three-point and four-point loading (Fig. 4). In three-point 
loading there is a significant shear stress generated at the midpoint of the beam. 
For this reason four-point loading is sometimes preferred because in a uniform 
specimen this is minimized and a pure bending moment is applied. However, in 
whole bones, the nonuniform cross section means that these assumptions do not 
apply. In addition, it is difficult to load all four points identically and test results 
will be subject to large errors. Three-point and four-point loading jigs can be 
bought or made to fit most materials testing machines. It is important to be 
able to adjust the span, the distance between the lower supports. For testing 
whole bones this means that the span may be matched to the length of the bone as 
this may differ between specimens. For machined specimens, testing at different 
spans enables extrapolation to be made for an infinite span (9). 

3.7. Tension Testing 

1. Prepare the specimen as described in Subheading 3.2. or 3.3. ensuring that it is 
at least 15-25 mm long and 5 mm across (see Note 4). 

2. Keep the sample moist by lightly wrapping in cling film or moist tissue or gauze. 

3. Fix the specimen in the testing machine. Most testing machines have grips for 
holding specimens in tension. These can be tightened on to the broader ends of 
the specimen and are adequate for testing cortical bone (see Note 5). If you are 

27/Aspden/367-380/F1 374 I 2/26/03, 10:49 AM 

I ■ I I I 

Mechanical Testing of Bone Ex Vivo 




Fig. 4. Whole bones may be tested in either (A) three-point or (B) four-point load- 
ing configuration. Force is applied through the upper plate and the resulting displace- 
ment recorded. The span is the distance between the lower supports. 

testing cancellous bone cores, set the ends into cups using dental or bone cement 
and the cups then held in the testing machine grips. 

4. Set up the machine to apply the desired strain rate, and program in any predeter- 
mined limits on load, displacement, or fracture detection, and the rate of data 
recording. It is advisable to set the machine to stop as soon as fracture occurs 
because at this point the load falls to zero. 

5. Start the testing machine. 

6. After the test, transfer data from the instrument to a spreadsheet for further analy- 
sis if the instrument does not already have appropriate software. 

3.8. Analyzing Data from Tension Testing 

1. If intrinsic properties are required, and not already calculated, convert load val- 
ues to stress by dividing by the original cross-sectional area. (No account is taken 
of changes in the cross-sectional area which will occur during the test.) 

2. Convert displacement values to strains by dividing by the unstretched distance 
between the points of attachment or reference marks of the displacement trans- 
ducer (see Note 5). 

3. Plot stress as a function of strain. The elastic, or Young's, modulus, yield strength 
and strain, ultimate strength and strain, and energies to yield and failure may then 
be calculated. 



2/26/03, 10:49 AM 

I ■ I I I 


376 Aspden 

4. Calculate the elastic modulus by estimating the gradient of the linear part of the 
curve. For linear relationships this is often calculated by the instrument software. 
If not, this can found crudely by hand calculation, better by fitting a straight line 
to the data or for nonlinear relationships, by differentiating the curve using soft- 
ware such as Origin. 

5. Calculate the yield, by estimating when the elastic modulus starts to reduce. This 
is not always easy to define, but one way is to construct a line parallel with the 
linear part of the curve but offset along the strain axis by a small amount. The 
point at which this intersects the curve is defined as the offset yield point and 
the stress and strain at this point are the yield stress and strain. 

6. Calculate the failure stress and strain from the point at which fracture occurs. 

7. Integrate the curve, to either the yield point or failure point, to find the area 
beneath to estimate the energy needed to produce yield or failure, respectively, 
per unit volume of material. The energy to failure is often referred to as the modu- 
lus of toughness. 

3.9. Compression Testing 

1. Prepare a cylinder or rectangular parallelepiped of bone as described in Sub- 
heading 3.2. Try to ensure that the length of the specimen is about twice the 
diameter, (see Note 6). 

2. Place the sample on the lower anvil of the testing machine. As in the case of 
tensile testing, a universal joint must be used in the loading path (see Note 4). 
Alternatively, a steel plate with small indentation containing a ball bearing 
between the specimen and the upper loading anvil can be used; this allows for 
any slight nonparallelism between the two faces. 

3. Accurate measurements of strain requires the use of a displacement transducer, 
as for tension, but for most comparative studies the displacement of the cross- 
head is probably adequate (see Note 5). 

4. Lower the crosshead of the machine until it is just in contact with the specimen. 

5. Set the strain rate and any predetermined load or displacement limits as before. 

6. Start the test and stop the test when the chosen criterion is met (see Note 7). 

7. Calculate stress and strain and plot stress as a function of strain as described in 
the preceding. Where testing is done between two anvils there is often a region 
near the origin where the slope of the curve starts small and increases rapidly 
before becoming approx linear, resulting in a J-shaped curve. This toe region is 
an artefact of the method and is commonly found in this sort of testing both of 
natural and synthetic polymeric materials. Because it can be difficult to deter- 
mine a straight portion of the curve or to fit a straight line to it, it is best to 
differentiate the curve to find the modulus as a function of strain. The elastic 
modulus can then be taken to be the peak value of this curve. 

8. The yield stress may be defined as the stress at which the modulus starts to 
decrease. We used a reduction of 3% from its peak value to provide a defined 
reference point (7), 

9. Energy to yield is found by integration of the area under the stress-strain curve 
up to the yield stress. Strains are ill defined in this testing protocol because of the 

27/Aspden/367-380/F1 376 1 2/26/03, 10:49 AM 

I ■ I I I 


Mechanical Testing of Bone Ex Vivo 377 

toe region. Embedding the ends of longer specimens and gripping as for a tensile 
test can eliminate this problem. Failure cannot usually be defined in the same 
way as for tensile testing because progressive crushing of the material often 
occurs rather than a fracture. 

3. 10. Bending 

Set the bending jig in the materials testing machine. 

1 . Adjust the spacing of the lower supports to be as close to the ends of the diaphy- 
sis as possible. 

2. Lower the crosshead of the instrument so that the upper loading anvil is just in 
contact with the bone surface. 

3. Set the displacement rate, any load or displacement limits, and the data collection rate. 

4. Start the test. 

5. The machine can be set to stop as soon as fracture occurs because at this point the 
load falls very rapidly, although not always to zero. 

6. Load is plotted as a function of displacement and the stiffness, yield load and 
failure load, and energy to fracture may be found using the same criteria as used 
for stress-strain curves in the tensile and compression tests (see Note 8). 

4. Notes 

1 . Storage of samples: Samples should be kept moist with phosphate-buffered saline 
(PBS) prior to mechanical testing; formalin and other fixatives must never be 
used as this affects the mechanical properties. If the samples are not being tested 
immediately (within 2-4 h), the specimens can be stored frozen at -20°C. They 
must be wrapped in tissue or gauze dampened with PBS and vacuum sealed into 
plastic freezer bags. A vacuum bag sealer can be obtained for this purpose. 

2. Instruments for precision cutting: We use an Accutom 2 (Struers Ltd, Glasgow) 
and a 125-mm aluminum oxide wheel, but the most common saw cited in the 
bone literature is the Isomet (Buehler, Lake Bluff, IL, USA). Whatever saw is 
used, faster wheel speeds and slower specimen feed results in a better surface 
finish. Where samples of a predetermined size are required, it is important to 
know how much of the sample will be lost during the cutting procedure. Any saw 
will remove an amount of material corresponding to the width of the blade, termed 
the kerf. For an aluminum oxide wheel the kerf is about 0.5 mm compared with 
about 0.3 mm for a diamond blade. When the kerf is known it is possible to 
obtain samples of the desired size by adjusting the cutting position appropriately. 
It is possible to cut specimens 100-150 [im thick using an aluminum oxide wheel. 
Care must be taken to keep specimen feed speeds, that is, the rate at which the 
specimen is moved past the cutting wheel, fairly low, as aluminum oxide wheels 
are brittle and shatter easily and diamond wheels are thin and are easily bent. 
However, aluminum oxide wheels cost about £10 ($15), in contrast to about 
£200-£300 ($200-300) for a diamond cutter, depending on size. 

3. Obtaining bone cores: When coring right through a block of tissue, the sample is 
commonly left in the coring bit. Occasionally this can happen, too, if a core snaps 

27/Aspden/367-380/F1 377 I 2/26/03, 10:49 AM 

I ■ I I I 


378 Aspden 

off while being drilled. To be able to remove this core we have had the bits made 
with a hole from end to end so that a close fitting plunger may be inserted to push 
out the core. However, this can damage the top of the core if it is tightly wedged 
into the drill bit. For this reason we prefer not to drill right through and to use the 
break-off tool to snap the core from its base after the coring bit has been re- 

4. Tension testing: The specimen must be large enough to get consistent results; 
Keaveny's methods start with a specimen 40 mm long, which highlights the dif- 
ficulties that may sometimes be encountered obtaining suitable specimens (6). If 
the grips are not exactly in line or the specimen is not precisely machined then 
applying a tensile load will also result in a degree of bending. This will add 
uncertainty to the test data which will be different for each specimen and make 
accurate comparison between specimens impossible. A universal joint must be 
placed in the path of the load to ensure that only tensile forces are transmitted to 
the specimen. These are available from the instrument manufacturers. 

5. Measuring strain during tension tests: An estimate of the strain can be calculated 
from the displacement of the crosshead recorded by the machine. However, for 
accurate studies a strain gauge or displacement transducer must be applied to the 
waisted part of the specimen. This can be done using either a clip-on device or 
video recording methods. 

6. Shape of samples for compression testing: It is desirable for the length of the 
specimen to be twice the diameter. It is possible to conduct tests on samples in 
which the dimensions are roughly equal without too much of a problem and in 
many cases, this may be unavoidable due to anatomical considerations. 

7. Setting limits for compression testing: Care needs to be taken when setting 
applied loads and limits into certain testing machines. Conventionally, engineers 
use positive numbers to denote tensile loads and negative numbers for compres- 
sive loads. The most commonly used test is tension and here there is generally no 
problem, as everything is positive. However, setting a limit for a compressive 
load often needs to include a negative sign. It sounds trivial but is easily over- 
looked and many students have stood watching the machine crush their specimen 
wondering why the machine has not stopped! Only after hitting the emergency 
stop button have they realized they omitted the negative sign in front of the load 
limit they had set. There is often no clear end point for failure in compression, 
unlike for tension, and determining when to stop the test has to be decided 
beforehand. In our studies we have decided that any reduction in stiffness — that 
is a reduction in the slope of the load-displacement curve — is a sign of the begin- 
ning of failure. As we usually wish to do further analysis on the specimens and 
minimize the damage caused by the test, we watch the load-displacement curve 
(which is plotted on the computer monitor during the test) carefully during the 
test, and stop the test as soon as it becomes obvious that the slope of the curve is 

8. Calculating intrinsic properties of bone from bending tests: Equations have been 
derived to calculate intrinsic properties from these data but are fraught with diffi- 

27/Aspden/367-380/F1 378 1 2/26/03, 10:49 AM 

I ■ I I I 


Mechanical Testing of Bone Ex Vivo 379 

culties when testing whole bones because of the asymmetric cross section of the 
bone. It is not recommended to use these without expert advice. 


1. Aspden, R. M. (1990) Constraining the lateral dimensions of uniaxially loaded 
materials increases the calculated strength and stiffness: application to muscle 
and bone. /. Mater. Sci. Mater. Med. 1, 100-104. 

2. Bryce, R., Aspden, R. M., and Wytch, R. (1995) Stiffening effects of cortical 
bone on vertebral cancellous bone in situ. Spine 20, 999-1003. 

3. Gibson, L. J. (1985) The mechanical behaviour of cancellous bone. /. Biomech. 
18, 317-328. 

4. Gibson, L. J. and Ashby, M. F. (1988) Cellular Solids. Pergamon Press, Oxford. 

5. Cowin, S .C. (2001) Bone Mechanics Handbook. CRC Press, Boca Raton, FL. 

6. Keaveny, T. M., Guo, X. E., Wachtel, E. F., McMahon, T. A., and Hayes, W. C. 
(1994) Trabecular bone exhibits fully linear elastic behavior and yields at low 
strains. /. Biomech. 27, 1127-1136. 

7. Li, B. and Aspden, R. M. (1997) Composition and mechanical properties of can- 
cellous bone from the femoral head of patients with osteoporosis or osteoarthritis. 
/. Bone Miner. Res. 12, 641-651. 

8. Turner, C. H. and Burr, D. B. (2001) Experimental techniques for bone mechan- 
ics, in Bone Mechanics Handbook (Cowin, S. C, ed.), CRC Press, Boca Raton, 
FL, pp. 7-1-7-35. 

9. Spatz, H. C, Oleary, E. J,, and Vincent, J. F. V. (1996) Young's moduli and shear 
moduli in cortical bone. Proc. R. Soc. B263, 287-294. 

27/Aspden/367-380/F1 379 I 2/26/03, 10:49 AM 

I ■ I I I 




2/26/03, 10:49 AM 

I ■ I I I 



Bone Cell Responses to Fluid Flow 

Hazel Y. Stevens and John A. Frangos 

1 . Introduction 

The relationship between mechanical loading and bone formation has long 
been documented (1). However, the identity of the transductory mechanism, 
conveying loading signals to bone cells, remains elusive. Mechanical strain, 
interstitial fluid flow, and streaming potentials are all likely candidates but the 
separate investigation of these factors has proven difficult. Studies in our labo- 
ratory and others have shown fluid shear stress to be influential in bone model- 
ing and remodeling (2-4). The characteristics of the flow causing this bone 
formation and probable inhibition of resorption have been more difficult to 

In the absence of loading, interstitial fluid, originating from leaky venous 
sinusoids in the intramedullary cavity, is driven radially outward through 
intracortical pores, the direction being dictated by a transmural pressure gradi- 
ent between the endosteal vasculature and the lymphatic drainage at the peri- 
osteal surface (5). Load-induced compression or bending of bone generates 
localized pressure gradients, which cause rapid fluid flow from areas of com- 
pression to areas where tension builds. After a loading event there is an associ- 
ated relaxation phase, hence a pulsatile flow ensues. Changes in interstitial 
flow may be responsible for the induction of bone formation in regions of 
elevated intraosseous pressure, for example, hypertension whereas normal 
pressures may serve to maintain normal bone architecture (6). 

The attainment of high fluid shear stresses within a short space of time, that 
is, temporal gradients in shear (TGS) are typically seen in high-impact exer- 
cise such as running and jumping (high amplitude but fairly infrequent strains) 
whereas posture maintenance involves a large number and high frequency of 
low-amplitude strains. Ramped flow is typified by the attainment of high shear 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


28/Frango/381-398/F1 381 1 2/26/03, 10:49 AM 

I ■ I I I 


382 Stevens and Frangos 

stress over a longer period of time. Given that impulse flow contains only tem- 
poral gradients and ramped flow involves only steady shear components with 
negligible temporal gradients, it is possible to subject cells to these different 
flow profiles and compare their responses. McAllister and Frangos (7) identi- 
fied two different mechanochemical transduction pathways in osteoblasts, 
depending on whether the stimulus was rapid onset of shear or steady shear. 
Preliminary data in our laboratory suggest that UMR-106 osteoblasts respond 
to transients in flow with increased cell proliferation and ERK phosphoryla- 
tion yet ramped flow appears inhibitory to ERK activation (8). 

This concurs with our studies on human endothelial cells (9). We have 
hypothesized that transients in shear mediate bone formation and deposi- 
tion by activating mitogenic factors whereas steady flow provides suffi- 
cient nitric oxide (NO) release to inhibit resorptive processes in localized 
bone remodeling. All bone cell types respond directly or indirectly to inter- 
stitial fluid flow. The signalling moieties include: inositol triphosphate 
(IP3; [10]; cyclic adenosine monophosphate (cAMP; [2]); transforming 
growth factor-(3 (TGF-(3; [11]); cellular fos proto-oncogene (c-fos [12]) and 
intracellular calcium in osteoblasts (13,14); prostaglandin E 2 (PGE 2 ) and 
prostacyclin (PGI 2 ) in osteoblasts and osteocytes (15) and nitric oxide (NO) 
in osteoblasts (3,16) and osteocytes (17). Preosteoclast-like cells are stimu- 
lated by fluid shear stress to release autocrine factors such as NO and PGE 2 
(18), which can regulate localized resorption in vivo (19). Many of these 
responses are G-protein-dependent (20) with the exception of steady shear 
production of NO (7) (see Fig. 1). 

The study of flow in vitro is complicated by the use of a range of different 
substrates, flow chambers, rates, characteristics of flow, as well as differences 
in preparation of cells (serum starvation, nutrients, confluency, etc). Although 
primary cells from mammalian donors are physiologically the most relevant 
cell types, they are difficult to isolate and characterize, therefore use of cell 
lines (e.g., UMR-106) has its place in such studies. In our laboratory, flow 
systems have been developed to test the responses of anchorage-dependent 
cells to a range of steady and pulsatile shear stresses under well-defined condi- 
tions. This chapter describes methods for using cultured cells to study effects 
of fluid flow in bone. 

2. Materials 

2.1. Tissue Culture Medium (UMR-106), Cell Line, and Substrate 

1. Dulbecco's modified Eagle's minimal essential medium (DMEM) with phenol 
red, ATP-free (Gibco, Grand Island, NY), supplemented to a final concentration 
of 10% with heat- inactivated fetal calf serum (FCS), 2 mmol/L of L-glutamine, 
50 U/mL of penicillin, and 50 |xg/mL of streptomycin for the maintenance of the 
UMR-106 rat osteosarcoma cell line. 

28/Frango/381-398/F1 382 I 2/26/03, 10:49 AM 

I ■ I I I 

Bone Cell Responses to Fluid Flow 








Fig. 1. Proposed scheme for G-protein signaling in the osteoblast. Fluid shear stress 
causes an increase in membrane fluidity leading to direct activation of G-proteins. NO 
production is G-protein and calcium insensitive under steady shear and is pertussis toxin 
(PTx)-insensitive G-protein mediated with rapid onset of shear (TGS). Prostaglandin E 2 
release is G-protein (partial PTx sensitive) mediated for both types of shear. a,p\y, sub- 
units of G-protein; Cav, caveolin; DAG, diacylglycerol; eNOS, endothelial nitric oxide 
synthase; ERK, extracellular signal regulated kinase; GDP, guanosine diphosphate; GTP, 
guanosine triphosphate; IP 3 , inositol triphosphate; MEK, ERK kinase; PGE 2 , prostag- 
landin E 2 ; PKC, protein kinase C; PLA 2 , phospholipase A 2 ; PLC, phospholipase C. 



2/26/03, 10:49 AM 

I ■ I I I 


384 Stevens and Frangos 

2. Serum-free DMEM supplemented with 0.1% bovine serum albumin (BSA) Frac- 
tion V cell culture tested (Sigma) and stored at 4°C; L-glutamine; penicillin; and 
streptomycin for the starving of monolayers before and during the flow experi- 

3. Hanks' balanced salt solution (HBSS) without calcium and magnesium (Gibco) 
for the washing of cell monolayers. 

4. Trypsin-EDTA (0.25% trypsin, 1 mM tetrasodiumethylenediaminetetraacetic 
acid (EDTA.4Na) for the detachment of cells from tissue culture flasks. 

5. Fibronectin-like engineered protein polymer ( 1 mL of diluent and 1 mg of powder 
combine to give a stock of 1 mg/mL; Sigma) stored at room temperature (RT). 

6. UMR-106 rat osteosarcoma osteoblast cell line (cat. no. CRL-1661, American 
Type Culture Collection [ATCC], Rockville, MD). 

2.2. Tissue Culture Medium (Primary Calvarial Osteoblast-Like 

1. M199 (Modified medium 199, with phenol red; Gibco) supplemented with 10% 
FCS, 1% penicillin-streptomycin, and 1% L-glutamine for the isolation and main- 
tenance of rat calvarial osteoblasts (complete medium). 

2. M199, ATP-free (custom made from Gibco), serum-free for the incubation of 
calvaria and the shear experiments. 

3. Fibronectin-like engineered polymer (see Subheading 2.1., item 5). 

4. 300-500-um Glass chips. 

5. Collagenase A (Roche Diagnostics, Berkeley, CA) stock, 1 mg/mL in HBSS 
stored in aliquots at -20°C. 

2.3. Equipment 

2.3. 1 Flow Chamber Apparatus 

The apparatus is shown diagrammatically in Figs. 2-4. 

1. The flow chamber (, La Jolla, CA) is a polycarbonate plate (Rohm 
and Haas, Philadelphia, PA), custom machine milled with a vacuum channel on 
one face. Laminar parallel flow is provided in a flow channel formed by a rectan- 
gular size 0.020-inch Silastic gasket (Dow Corning, Midland, MI) and a glass 
slide with adherent cell monolayer. Standard luer female fittings are at fluid inlet/ 
outlet and vacuum ports. Interconnecting tubing (1/8-inch inner diameter) of 
polytetrafluoroethylene (PTFE, Teflon) (Cole Parmer, Chicago, IL) is fitted with 
luer male connectors. The relatively inert and gas impermeable tubing prevents 
water and gas loss and minimizes absorption of metabolites. 

2. Syringe pump (pump 44, Harvard Apparatus, Holliston, MA), syringes (100-mL 
gas-tight borosilicate glass with PTFE [Teflon] plunger and luer lock; Scientific 
Glass Engineering Inc, Austin, TX and 30-mL glass syringe; Becton Dickinson, 
Franklin Lakes, NJ). 

3. 37 C C Air box, with a high-throughput blower to provide heating and a thermo- 
static regulator. 

28/Frango/381-398/F1 384 I 2/26/03, 10:49 AM 

I ■ I I I 

Bone Cell Responses to Fluid Flow 


Fig. 2. Flow chamber apparatus for short-term experiments: A, computer-controlled 
syringe pump; B, syringe; C, flow chamber; and D, collector of perfusate. 


Fig. 3. Parallel plate flow chamber. The polycarbonate plate, the gasket (G), and 
the glass slide (H) with the attached cells are held together by a vacuum (C), forming 
a channel of parallel plate geometry. Medium enters at entry port (A), through slit (£), 
into the channel, and exits through slit (F), and exit port (B). Entry port (A) also serves 
as a trap for bubbles, which can be removed through valve (D). (From: Frangos, J. A., 
Mclntire, L. V., and Eskin, S. G. (1988) Shear stress induced stimulation of mamma- 
lian cell metabolism. Biotechnol. Bioeng. 32, 1053-1060. Copyright ©1988. Reprinted 
by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.) 



2/26/03, 10:49 AM 

I ■ I I I 


Stevens and Frangos 


Fig. 4. Pulsatile flow loop. An offset cam (A) driven by an electric motor drives 
syringe plunger (B) through one-way check valves (C), forcing flow through chamber 
(D), recirculating into the reservoir. Media samples may be drawn aseptically from 
valve (E) (Hillsley MS Thesis 1990). (From Hillsley et al., Calcified Tissue 
International 1997; 60, 48-53, with permission of Springer- Verlag New York, Inc.) 

4. 38 x 75 mm Glass microscope slides, thickness 0.90-1.10 mm (Corning Inc, 
Acton, MA). 

5. Offset cam driven by a stepper motor (Anaheim Automation, Anaheim, CA; 23A 
108C with smc40x driver/indexer) and syringe. 

2.3.2. Glass Reservoir for Fluid Loop 

1. One-way check valves (Becton Dickinson). 

2. Optional electromagnetic flow probe (Zepeda Instruments, Seattle, WA). 

2.3.3. Flow Loop (see Fig. 5) 

1. Upper and lower glass reservoirs (, roller pump (Becton 
Dickinson). PTFE tubing (0.125-inch outer diameter, Cole Parmer) except for 
section through roller pump which is silicone (Masterflex, Cole Parmer). Sili- 
cone collars join the reservoirs to the manifold and tubing. 

3. Overflow glass manifold (19-mm outer diameter, Pyrex, Corning Glass Works, 
Corning, NY). 

2.4. Cell Assays 

1. Dextran (2,000,000 mol wt; Sigma). 

2. Chemiluminescent flow analyzer, Radical Purger (ASM 03296) and Nitrate 
Chemical Kit (ASK 14400-01) (Sievers Instruments, Boulder, CO). 

3. Bradford protein assay (Bio-Rad, Hercules, CA). 



2/26/03, 10:49 AM 

I ■ I I I 

Bone Cell Responses to Fluid Flow 



Fig. 5. Drawing of small volume flow loop: (1) upper reservoir, (2) lower reservoir, 
(3) overflow manifold, (4) filtered humidified 95% air + 5% C0 2 input, (5) gas outlet, 
(6) flow chamber, (7) gasket, (8) slide with cell monolayer, (9) microscope objective, 
(10) vacuum, (11) sampling port, (12) roller pump, (13) PFA teflon tubing, (14) con- 
stant pressure head, and (15) flow probe. (From: Frangos, J. A., Mclntire, L. V., and 
Eskin, S. G. (1988) Shear stress induced stimulation of mammalian cell metabolism. 
Biotechnol. Bioeng. 32, 1053-1060. Copyright (c) 1988. Reprinted by permission of 
Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.) 

4. Aminoguanidine (AG; Alexis Biochemicals, San Diego, CA) and N G -amino l- 
arginine (l-NAA; Alexis). 

5. Alkaline phosphatase 86-C kit (Sigma). 

6. GDP(3S, GTPyS (Calbiochem), pertussis toxin (Sigma). 



2/26/03, 10:49 AM 

I ■ I I I 


388 Stevens and Frangos 

3. Methods 

3. 1. Preparation of UMR-106 Cells for Flow Experiments 

1. Prepare rectangular glass microscope slides (38 x 75 mm) for cell seeding by 
fibronectin coating. Place autoclaved slides into 100-mm Petri dishes and dilute 
fibronectin polymer stock to 10 ng/mL with sterile phosphate-buffered saline 
(PBS; Irvine Scientific, Santa Ana, CA) before using 1 mL to coat each slide. 
Leave the slides for 5 min at RT before removing the excess, rinsing twice in 
PBS and leaving the slides to air-dry in the flow hood (class II laminar flow 
hood) for at least 1 h. 

2. Aspirate medium from a 75-cm 2 flask of cells and wash the cells in HBSS. Add 1 mL 
of trypsin and incubate the flask for approx 10 min at 37°C, after which time the 
cells should have become detached. Add 9 mL of DMEM containing 10% FCS to 
the flask in stages and pool the washes. Introduce 20 (iL into a hemocytometer to 
determine cell count and centrifuge the rest at 240g for 5 min. 

3 . Resuspend the cells in sufficient DMEM with 10% FCS for seeding 0.7 mL of cells onto 
the slides at 0.8 x 10 6 cells/mL and let them grow to confluence for 2-3 d in a humidified 
incubator at 5% C0 2 -95% air at 37°C. Before exposure to shear, place a negative tem- 
plate of the gasket underneath each Petri dish and scrape the cells from this region to avoid 
the inclusion of cells squashed by the gasket. Serum starve the confluent cell monolayers 
for 24 h in 0. 1 % BS A-DMEM before subjecting them to flow. Dissolve the BS A powder 
in medium and then sterilize by passing it through a 0.2-\im filter. 

3.2. Preparation of Newborn Rat Calvarial Osteoblast-Like Cells 
for Flow Experiments 

This method employs culture techniques adapted from Ecarot-Charrier et al. 
(21), the premise being the selective migration of osteoblasts onto glass chips 
(22). The protocol, described in detail by Reich et al. (2), uses the calvaria 
from 3-6-d-old Sprague-Dawley rat pups. 

1. Remove the calvaria aseptically and place them into Ml 99 with no additives. 
Strip the periosteum carefully from both sides using watchmakers' forceps and 
place the bone into Petri dishes. 

2. Place glass chips on the endocranial surface of the calvaria and add complete medium. 

3. After 4-6 d, dislodge the glass chips from the calvaria by irrigation with com- 
plete medium and incubate the glass chips for an additional 7-10 d before use, 
replacing the medium every 3—4 d. 

4. Remove the cells from the glass chips by collagenase digestion (30 min, 0.2 mg/mL 
in HBSS) and resuspend the cells in complete medium. Allow cells to grow until 
confluent in culture flasks. 

5. Trypsinize the cells as in Subheading 3.1. and seed 1 mL of the cells onto glass 
slides coated with fibronectin until they reach confluence, in complete medium. 
Differentiation of osteoblastic precursor cells is encouraged by the addition of 
50 u.g/mL of ascorbic acid and 5 mM |3-glycerophosphate to the medium. The 
cells are characterized morphologically and functionally by their alkaline phos- 

28/Frango/381-398/F1 388 1 2/26/03, 10:49 AM 

I ■ I I I 


Bone Cell Responses to Fluid Flow 389 

phatase (AP) activity and their ability to form mineralized nodules (see Note 1). 

3.3. Flow Experiments (see Note 2) 

3.3.1. Apparatus for Computer Driven Syringe Pump Model (Short- 
Term Experiments) 

The entire flow device is housed in a 37°C air box (see Fig 2.). The fluid 
used to shear the cells is the same as the incubation medium prior to shear. The 
parallel plate flow chamber is perfused by a syringe pump. The channel base 
area is 13 cm 2 and the channel depth is nominally 230 ^im (see Fig. 3). Perfu- 
sion medium is equilibrated with 5% C0 2 , 37°C and the apparatus is allowed 
to equilibrate to 37°C. 

1 . Fill 100-mL and 30-mL glass syringes with 37°C medium and expel any bubbles (see 
Note 3). Prime all the tubing, including the exit slit and port. Introduce medium into 
the chamber by means of the pump so that the medium fills the area to be occupied by 
the slide. Invert the first slide over the flooded chamber by gently placing one edge 
near the inlet port and lowering it gently to avoid any air bubbles. Once the slide is 
positioned on the chamber, apply uniform pressure and attach the vacuum line to the 
vacuum port to keep the slide in place and to ensure uniform channel depth. Excess 
medium will flow out so it is advisable to cover the chamber support with absorbent 
material. The outlet port tubing can be placed into a beaker to collect perfusate as 
long as all tubes entering and exiting the chamber are maintained at the same eleva- 
tion. At all times the chamber must be handled carefully to minimize shear and to 
avoid temporal gradients in particular. The entry port is larger than the exit port to 
serve as a bubble trap and once the vacuum is applied, the whole chamber can be 
inverted and bubbles removed via the septum port. Trapped air bubbles can interfere 
with the fluid flow and change its characteristics (see Note 3). 

2. The syringe pump is programmed to generate a range of different flow profiles, 
the flow rate (see Note 4) being changed in 80-ms microsteps. The fluid dynam- 
ics of the chamber are described in greater detail in (23) (see also Note 4). Four 
well-defined laminar flow profiles are used to separate the effects of different 
flow stimuli in bone cells (see Fig. 6). 

a. Stepped flow (instantaneous [100 ms] shear stress increase from to 16 dyn/cm 2 
followed by steady shear for 10 min). 

b. Ramp flow (shear stress smoothly transitioned from to 16 dyn/cm 2 over 
30 sec, sustained for 9 min before ramping down over 30 sec). 

c. Impulse flow (a 500-ms to 3-sec impulse of 16 dyn/cm 2 with cells maintained 
on the flow chamber at 0.016 dyn/cm 2 for 10 min). 

d. Pulsatile flow (multiple flow impulses with 3-sec intervals). Slides are main- 
tained on the flow chamber for the same overall amount of time. The inclu- 
sion of time-matched stationary (in Petri dishes) and sham controls (placed 
on the chamber but not subjected to shear, only the vacuum) is important in 
defining the response of cells to fluid flow alone. 

28/Frango/381-398/F1 389 1 2/26/03, 10:49 AM 

I ■ I I I 


Stevens and Frangos 

1. Stepped flow 

2. Ramped flow 


steady flow 




600,000 ms 

ms Time 10 mins 

100ms Tlme 


3. Impulse flow 

4. Pulsatile flow 






500 ms 


10 mins 




10 mins 

3000 ms 
Fig. 6. Flow profiles for the experiments detailed in Subheading 3.3. 

3.3.2. Apparatus for Pulsatile Flow Loop Model 

Longer term studies on pulsatile flow require the apparatus shown in Fig. 4. 

1. Fill the syringe and release any trapped air bubbles. Fill the tubing until there is 
no air in the system. 

2. Attach the syringe to the offset cam and motor. The syringe will draw in medium 
from the reservoir through a one-way valve and force it out through another one- 
way check valve to the chamber. 

3. Adjust the flow rate by changing the frequency of pulsation of the syringe and 
the stroke volume. Experiments can be run for 15 min to 24 h (see Note 5). 

3.3.3. Apparatus for Steady Flow Loop Model 

Longer term steady flow studies on bone cells require the apparatus shown 
in Fig. 5. The flow loop consists of two reservoirs, situated one above the other 
with the flow chamber between them. Flow is maintained by the hydrostatic 
pressure head created by the vertical distance between the two reservoirs. The 
continuous pumping of the culture medium from the lower to the upper reser- 



2/26/03, 10:49 AM 

I ■ I I I 


Bone Cell Responses to Fluid Flow 391 

voir, at rates greater than those through of the chamber, maintains the pressure 
head. As the excess fluid drains down the glass overflow manifold it meets the 
incoming 5%C0 2 -95% humidified air, which facilitates gas exchange with the 

1. Add medium to the top reservoir (10-20 mL), filling the lower reservoir as well 
and flooding the chamber. 

2. Add the slide as before and adjust the fluid shear stress by changing the column 
height in the loop. Recirculation allows for the quantitation of cumulative NO 
production. In the PGE 2 studies, it is advisable to add serum to the medium to 
ensure that arachidonic acids, necessary for the response, are not rate limiting. 
Using this model Johnson et al. (16) demonstrated the rapid and continuous re- 
lease of NO in primary rat calvarial osteoblasts. 

3.4. Assays (see Note 6) 

The practical applications of these fluid flow models are numerous, as they 
allow for in situ studies of perfusate and cells as well as production of lysates 
and immunocytochemistry. A popular application in our laboratory has been 
their use to investigate signaling mechanisms in fluid shear. A brief descrip- 
tion of some of these techniques, to study G-protein-regulated pathways, fol- 

3.4.1. Viscosity and Shear Rate Experiments 

The role of viscosity is investigated by addition of high molecular weight, 
neutral dextran to the perfusion medium (2 g/100 mL, viscosity p = 3.9 mPa-s. 
[The viscosity of water at 20°C =1 mPa-s]). Normal viscosity of the medium is 
therefore assumed to be 1 mPa-s = 0.01 dyn s/cm 2 . Using the equation for 
viscosity (see Note 4), the viscosity and shear rate can be varied independently. 
Three or more profiles can be used: 

1. Wall shear stress (Tj) =26 dyn/cm 2 where viscosity (jx) = 0.01 and shear rate (y) = 
2600 s" 1 . 

2. x 2 = 26 dyn/cm 2 , \i = 0.039 (dextran added) and y = 660 s _1 . 

3. x 3 =6 dyn/cm 2 , p= 0.01, and v = 600 s 1 . 

In calvarial osteoblasts the rate of NO production increases with shear stress 
magnitude (0.25-26 dyn/cm 2 ) within the postulated physiological shear stress 
range (Fig. 7) (7). For a given viscosity (with wall shear stress r l and x 3 ) the 
increase in shear rate is associated with an increase in NO release and for simi- 
lar shear rates (wall shear stress x 2 and x 3 ) the increased viscosity stimulates 
increased NO production (Fig. 8) (7,24). At equal flow rate the streaming 
potentials and mass transport are not affected by increasing the viscosity but 
the wall shear stress is directly proportional to viscosity. Therefore wall shear 
stress is the main mechanotransductory factor in NO production. 

28/Frango/381-398/F1 391 1 2/26/03, 10:49 AM 

I ■ I I I 


Stevens and Frangos 










Shear Stress (dyne/cm 2 ) 

Fig. 7. NO production rates (0-6 h) in primary rat calvarial osteoblasts demonstrate 
a dose response across the range of physiological shear stress. (Reproduced from /. 
Bone Miner. Res. 1999; 14, 930-936 with permission of the American Society for 
Bone and Mineral Research.) 








fc 1 L 2 l 3 

Fig. 8. Viscosity (\i) and shear rate (y) were varied independently to investigate the 
role of viscosity and wall shear stress (x). NO production rates (0-6 h) are plotted for 
three cases: x x = 26 dynes/cm 2 where \i = 1.0 and y = 2600; x 2 =26 dynes/cm 2 where [i = 
3.9 and y = 660; x 3 = 6 dynes/cm 2 where [i = 1.0 and y = 600. f x : and x 2 were signifi- 
cantly different from x 3 (p < 0.05). (Reproduced from /. Bone Miner. Res. 1999; 14, 
930-936, with permission of the American Society for Bone and Mineral Research.) 

3.4.2. NO Release Measurement 

1. Draw 500-p.L samples from the perfusate (from the outflow in the computerized 
flow apparatus and from the valve at the base of the reservoir in the flow loop 



2/26/03, 10:49 AM 

I ■ I I I 


Bone Cell Responses to Fluid Flow 393 

2. Reduce the nitrite and nitrate in the sample to NO using a vanadium chloride 
reaction vessel (Radical Purger and Nitrate Chemical Kit [Sievers]). In brief, 
vanadium (III) chloride in hydrochloric acid converts nitrate to NO and it also 
converts nitrite and 5-nitroso compounds to NO. The reaction takes place at 90°C 
for high conversion efficiency with 5 mL of reducing agent for 20-50 samples. 

3. Quantify NO content using a chemiluminescent NO analyzer. Samples are com- 
pared with standard curves, then normalized to total cellular protein using the 
Bradford protein assay. Controls without cells are included to account for nitrates 
leeching from the loop apparatus or sampling equipment. 

3.4.3. Measurement of NO Synthase Inhibition 

1. Conduct flow studies with UMR-106 cells pretreated with the nitric oxide syn- 
thase (NOS) inhibitors aminoguanidine (AG ) at 100 \iM and l-NAA at 100 \M 
prior to flow and during flow, for example, for 6 h at 12 dyn/cm 2 . 

2. l-NAA is stable at RT but has poor solubility in water so a stock in HC1 is diluted 
further in culture medium. 

3. AG is a selective inhibitor of iNOS (NOS II), and l-NAA inhibits both constitu- 
tive and inducible NOS. 

4. In UMR-106 cells the iNOS inhibitor does not attenuate NO production whereas 
the general inhibitor of NOS significantly inhibits the flow-induced response 
(Fig. 9) (7). The form of NOS activated by shear is constitutive and is likely to be 
NOS III (eNOS) demonstrated to be present in osteoblasts (25). 

3.4.4. The Role of G-Proteins in NO Production 

1. Conduct flow studies with UMR-106 cells and primary calvarial osteoblasts 
preincubated with G-protein inhibitors GDP|3S (300-900 \iM) for 2 h, pertussis 
toxin 1 (xg/mL, 1 h. 

2. Incubate static cultures of osteoblasts on glass slides with or without G-protein 
activator GTPyS (300-900 \iM) for 2 h. 

3. NO production with flow is biphasic, with a point of inflection at 30 min. The 
rate of production of NO from to 30 min is about fourfold that of the sustained 
response. Static cultures treated with GTPyS showed a dose dependent increase 
in NO production rates. In sheared cultures the inhibitor GDP|3S blocks the ini- 
tial burst of NO but does not affect the sustained production (Fig. 10). Pertussis 
toxin, inhibitory for GJG class of G-proteins did not significantly affect either 
phase of NO production. 

4. Notes 

1. For AP staining follow the manufacturer's instructions (Sigma kit 86-C). In brief, 
the kit comprises a naphthol AS-BI phosphate alkaline solution and a Fast Blue 
BB salt. The AP enzyme cleaves the phosphate from the ASBI salt, which in turn 
reacts with Fast Blue to give a blue pigment. Discrete granular deposits of pig- 
ment can be seen microscopically and reflect areas of AP activity. The primary 
osteoblast-like cells stain for AP activity, usually in patches. Cells grown in the 

28/Frango/381-398/F1 393 1 2/26/03, 10:49 AM 

I ■ I I I 


Stevens and Frangos 












Static L-NAA AG 


Fig. 9. Cumulative NO production for flow at 12 dyn/cm 2 , flow + AG, flow + l- 
NAA, and static controls in UMR-106 cells. *l-NAA significantly inhibited NO re- 
lease (p < 0.05). (Reproduced from /. Bone Miner. Res. 1999; 14, 930-936, with 
permission of the American Society for Bone and Mineral Research.) 


Fig. 10. Cumulative NO production for flow at 16 dyn/cm 2 (filled circle), flow + 
GDP|3S (filled diamond), flow + PTx (open square), and static controls (open triangle) 
in primary rat calvarial osteoblasts. Inset: Short- and long-term production rates are 
shown for flow (horizontal hatching), flow + PTx (diagonal hatching), and flow + 



2/26/03, 10:49 AM 

I ■ I I I 


Bone Cell Responses to Fluid Flow 395 

presence of ascorbic acid and ^-glycerophosphate form mineralized nodules as 
determined by von Kossa staining (see the chapter by Scutt et al., this volume, for 
details on procedure). The cells show a threefold increase in cAMP in response to 
15 min parathyroid hormone stimulation, which is characteristic of osteoblasts 
and they produce osteocalcin (2). In our experience cell viability, assessed by 
trypan blue staining, is >98% in all experiments. 

2. Studies on cell metabolism usually require minimal flow volume (loop fluid) and 
maximum cell surface area, and the design of this flow chamber gives a high cell 
to volume ratio (10 6 cells/10 mL), suitable for the study of metabolite produc- 
tion. The slide area is also sufficient for the collection of lysates and the process- 
ing of samples for immunoblotting. A flow chamber with slightly reduced 
proportions can be mounted on an inverted microscope, allowing for time lapse 
photography of the cells. Several groups have used a cone and plate viscometer, 
which gives a smaller cell to volume ratio, does not allow for continuous sam- 
pling of the perfusate, and suffers from substantial evaporation of the tissue cul- 
ture medium. Some alternative fluid flow loops require large amounts of medium 
(in excess of 100 mL) for perfusion and have a relatively small cell surface area 
(1 cm 2 ); hence they are not so reliable for metabolic studies. 

3. In our experience, the computer-controlled syringe pump is essential for the 
delivery of rapid yet precisely controlled bursts of flow. The cells, if grown to 
confluence on coated/alkali etched slides, do not lose adherence during the flow 
cycle. However, a stray air bubble in the flow loop can easily tear cells from the 
slide and occasionally, on application of the vacuum, there is an air leak between 
the gasket and the chamber which will lead to a much greater volume of fluid 
being evacuated from the slide. To prevent these problems it is important to check 
for air bubbles before each flow experiment and to pass medium through to 
release any bubbles. In addition, the silastic gasket must be fully flush with the 
chamber and sealed with a little vacuum grease. It is advisable to set up a few 
extra slides in case one of the slides fails owing to these problems. Glass syringes 
are superior to plastic syringes as they respond more quickly to the action of the 
syringe pump owing to low compliance and are necessary for the correct delivery 
of short impulse flow. The chamber should be protected from scratches and chips 
as this will affect the flow characteristics. The length of the outlet should be kept 
to a minimum to avoid resistance to flow and at the same elevation as the cham- 
ber outlet to avoid suction or back pressures. 

4. The shear stresses of 0-16 dyn/cm 2 (1 dyn/cm 2 = 0. 1 Pa) are calculated from the 
velocity of flow, medium viscosity and dimensions of the chamber. The chamber 
is engineered to give flow rate and shear stress a linear relationship. The wall 
shear stress on the cell monolayer can be calculated using the momentum balance 
for a Newtonian fluid and assuming parallel plate geometry: 

Fig. 10. (continued) GDP|3S (white). GDP(3S significantly inhibited short-term NO 
production (p < 0.05). (Reproduced from/. Bone Miner. Res. 1999; 14, 930-936, with 
permission of the American Society for Bone and Mineral Research.) 

28/Frango/381-398/F1 395 1 2/26/03, 10:49 AM 

I ■ I I I 

396 Stevens and Frangos 

x = (6Qn)/(bh 2 ) 

where Q is the flow rate (mL/s); \i is the viscosity (~ 0.01 dyn s/cm 2 at 25°C); h is 
the channel height (0.023 cm); b is the slit width (2.35 cm); and x is the wall shear 
stress (dyn/cm 2 ). Therefore for a shear stress of 16 dyn/cm 2 the flow rate in mL/ 
sec is: Q = 0.332 mL/s = 19.89 mL/min for one syringe (9.94 mL/min for two 

The Reynolds number for flow through the chamber is 0-20 for shear stresses up 
to 24 dyn/cm. 2 The transition to turbulent flow starts at 1000-8000 Reynolds so 
the flow can be considered laminar. The entrance length for this flow chamber is 
0.018 cm (length required before the channel flow is truly laminar), and as the 
chamber is in effect approx 6.4 cm in length, the region at the top of the chamber 
where flow is nonlaminar is negligible. 
The wall shear stress (x) in dynes per square centimeter is defined as : 

x = y\i 

where y = shear rate s _1 and u, = dynamic viscosity (dyn s/cm 2 ). 

5. Flow can be monitored by the introduction of a flow sensor to the system. The 
flow pattern generated by this system is essentially a half sine wave. One cycle of 
the cam constitutes a half-sinusoidal positive flow and then a period of no flow. 
The peak flow rate and therefore peak shear stress is Jt x the average flow rate, 

-{©)- the average flow rate being calculated through an entire cycle. As shear stresses 

on osteocytes are calculated to be in the range 8-30 dyn/cm 2 (26), an average 
pulsatile shear stress of 5 dyn/cm 2 is used (giving a maximum shear stress of 5 x 
it =15 dyn/cm 2 ). 

6. Most experimental time is taken up with the passaging of cells and preparation of 
slides ready for the flow experiments. Depending on the conditions used, the 
flow experiments are relatively short and the most time is expended in heating 
the medium, air hood, and apparatus to 37°C. The perfusion medium can be left 
overnight gassing in C0 2 , as it is important to maintain the correct pH during the 
flow study and the medium rapidly loses C0 2 in the air hood. Fresh medium can 
be taken from the closed gassed bottle for each new experiment. Owing to the 
biohazardous nature of the waste from this experiment all syringes, plastic tub- 
ing, and flow chamber are washed in Tergazyme (Alconox Inc., New York, NY) 
and then rinsed in distilled H 2 before air drying. The silastic gasket does not 
retain its shape after autoclaving but can be subjected to ultraviolet light in a flow 
hood. For short term experiments sterility is not an issue. However, for long-term 
experiments it is advisable to autoclave the respective parts of the apparatus and 
assemble them in a laminar flow hood. 


The authors wish to thank Dr. Mark Haidekker for critical reading of the 

28/Frango/381-398/F1 396 1 2/26/03, 10:49 AM 

I ■ I I I 


Bone Cell Responses to Fluid Flow 397 


1. Wolff, J. (1892) Das Gesetz der Transformation der Knochen. Hirschwald, Berlin. 

2. Reich, K. M., Gay, C. V., and Frangos, J. A. (1990) Fluid shear stress as a media- 
tor of osteoblast cyclic adenosine monophosphate production. /. Cell. Physiol. 
143, 100-104. 

3. Smalt, R., Mitchell, F. T., Howard, R. L„ and Chambers, T. J. (1997) Induction of 
NO and prostaglandin E2 in osteoblasts by wall-shear stress but not mechanical 
strain. Am. J. Physiol. 273, E751-E758. 

4. Bergula, A. P., Huang, W., and Frangos, J. A. (1999) Femoral vein ligation in- 
creases bone mass in the hindlimb suspended rat. Bone 24, 171-177. 

5 . McAllister, T. N. and Frangos, J. A. (1998) Nitric oxide and mechanical factors: fluid shear 
stress, in Nitric Oxide in Arthritis and Osteoporosis (Hukkanen, M. V. J., Polak, J. M., and 
Hughes, S. P. F., eds.), Cambridge University Press, Cambridge, UK, pp. 141-150. 

6. Hillsley, M. V. and Frangos, J. A. (1994) Review: bone tissue engineering: The 
role of interstitial fluid flow. Biotechnol. Bioeng. 43, 573-581. 

7. McAllister, T. N. and Frangos, J. A. (1999) Steady and transient fluid shear stress 
stimulate NO release in osteoblasts through distinct biochemical pathways. /. 
Bone Miner. Res. 14, 930-936. 

8. Jiang, G. L„ White, C. R., Stevens, H. Y., and Frangos, J. A. (2002) Temporal 
gradients in shear stimulate oseoblastic proliferation via ERK1/2 and retinoblas- 
toma protein. Am. J. Physiol. Endocrinol. Metab. 283, E383-E389. 

9. Bao, X., Clark, C. B., and Frangos, J. A. (2000) Temporal gradient in shear- 
induced signaling pathway: involvement of MAP kinase, c-fos, and connexin 43. 
Am. J. Physiol. Heart Circulat. Physiol. 278, H1598-H1605. 

10. Reich, K. M. and Frangos, J. A. (1991) Effect of flow on prostaglandin E2 and 
inositol trisphosphate levels in osteoblasts. Am. J. Physiol. 261, C428-C432. 

11. Sakai, K., Mohtai, M., and Iwamoto, Y. (1998) Fluid shear stress increases trans- 
forming growth factor beta 1 expression in human osteoblast-like cells: modula- 
tion by cation channel blockades. Calcif. Tissue Int. 63, 515-520. 

12. Pavalko, F. M., Chen, N. X., Turner, C. H., et al. (1998) Fluid shear-induced 
mechanical signaling in MC3T3-E1 osteoblasts requires cytoskeleton-integrin 
interactions. Am. J. Physiol. 275, C1591-C1601. 

13. Hung, C. T„ Pollack, S. R., Reilly, T. M., and Brighton, C. T. (1995) Real-time 
calcium response of cultured bone cells to fluid flow. Clin. Orthopaed. Relat. Res. 
313, 256-259. 

14. McDonald, F., Somasundaram, B., McCann, T. J., Mason, W. T., and Meikle, M. 
C. (1996) Calcium waves in fluid flow stimulated osteoblasts are G protein medi- 
ated. Arch. Biochem. Biophys. 326, 31-38. 

15. Klein-Nulend, J., van der Plas, A., Semeins, C. M., et al. (1995) Sensitivity of 
osteocytes to biomechanical stress in vitro. FASEB J. 9, 441-445. 

16. Johnson, D. L., McAllister, T. N., and Frangos, J. A. (1996) Fluid flow stimulates 
rapid and continuous release of nitric oxide in osteoblasts. Am. J. Physiol. 271, 

28/Frango/381-398/F1 397 I 2/26/03, 10:49 AM 

I ■ I I I 


398 Stevens and Frangos 

17. Klein-Nulend, J., Semeins, C. M., Ajubi, N. E., Nijweide, P. J., and Burger, E. H. 
(1995) Pulsating fluid flow increases nitric oxide (NO) synthesis by osteocytes 
but not periosteal fibroblasts — correlation with prostaglandin upregulation. 
Biochem. Biophys. Res. Commun. 217, 640-648. 

18. McAllister, T. N., Du, T., and Frangos, J. A. (2000) Fluid shear stress stimulates 
prostaglandin and nitric oxide release in bone marrow-derived preosteoclast-like 
cells. Biochem. Biophys. Res. Commun. 270, 643-8. 

19. Kasten, T. P., Collin Osdoby, P., Patel, N., et al. (1994) Potentiation of osteoclast 
bone-resorption activity by inhibition of nitric oxide synthase. Proc. Natl. Acad. 
Sci. USA 91, 3569-3573. 

20. Reich, K. M., McAllister, T. N., Gudi, S., and Frangos, J. A. (1997) Activation of 
G proteins mediates flow-induced prostaglandin E2 production in osteoblasts. 
Endocrinology 138, 1014-1018. 

21. Ecarot-Charrier, B., Glorieux, F. H., van der Rest, M., and Pereira, G. (1983) 
Osteoblasts isolated from mouse calvaria initiate matrix mineralization in culture. 
J. Cell Biol. 96,639-643. 

22. Jones, S. J. and Boyde, A. (1977) The migration of osteoblasts. Cell Tissue Res. 
184, 179-193. 

23. Frangos, J. A., Mclntire, L. V., and Eskin, S. G. (1988) Shear stress induced stimu- 
lation of mammalian cell metabolism. Biotechnol. Bioeng. 32, 1053-1060. 

24. Bakker, A. D., Soejima, K., Klein-Nulend, J., and Burger, E. H. (2001) The pro- 
duction of nitric oxide and prostaglandin E(2) by primary bone cells is shear stress 
dependent. /. Biomech. 34, 671-677 '. 

25. Helfrich, M. H., Evans, D. E., Grabowski, P. S., Pollock, J. S., Ohshima, H., and 
Ralston, S. H. (1997) Expression of nitric oxide synthase isoforms in bone and 
bone cell cultures. /. Bone Miner. Res. 12, 1 108-1 115. 

26. Weinbaum, S., Cowin, S. C, and Zeng, Y. (1994) A model for the excitation of 
osteocytes by mechanical loading-induced bone fluid shear stresses. /. Biomech. 
27, 339-360. 

28/Frango/381-398/F1 398 1 2/26/03, 10:49 AM 

I ■ I I I 



Methods for Analyzing Bone Cell Responses 
to Mechanical Loading Using In Vitro Monolayer 
and Organ Culture Models 

Andrew A. Pitsillides, Victoria Das-Gupta, Dominic Simon, 
and Simon C. F. Rawlinson 

1. Introduction 

As bone's primary function is mechanical, it is not surprising that almost all 
studies using intact bone concern its morphology. Such histomorphometric 
studies have been used to provide insights into how bone responds, as an organ, 
to mechanical loading. However, despite the fact that the cellular basis for 
"sensing" mechanical stimuli or "communicating" their influence to coordi- 
nate any loading-induced changes that they engender is not known, studies in 
intact bone are rarely used to establish the direct links with any changes in 
bone cell biochemistry. It is also evident that most studies aimed at defining 
these mechanisms currently use bone cells grown in vitro, and that this has 
produced rapid advances in our understanding of the factors that might be 
involved in regulating bone cell responses to loading-induced stimuli. It is clear 
that such in vitro studies facilitate the final mechanistic deciphering and con- 
stitute a useful initial approach. However, it is also evident that they generally 
take little regard of the influence that might be provided by cell-cell and cell- 
matrix interactions within a bone's complex environment and architecture (1). 
It is therefore imperative to attempt to bridge the gap between the cell biology 
of bone's response to loading on the one hand and the morphological approach 
to this same problem on the other. 

In this chapter, we concentrate on techniques by which the cellular responses 
to mechanical stimulation can be examined in cell monolayer and in organ 
culture. We provide details on a wide range of devices that can be used to 
generate precise, measurable mechanical strains in these preparations and give 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


29/Pitsillides/399-422/F1 399 1 2/26/03, 10:56 AM 

I ■ I I I 

400 Pitsillides et al. 

details of those devices that we have used extensively in both organ and cell 
culture. We hope that this provides a basis upon which new investigators into 
this field can decide how to approach their own particular set of questions. In 
our opinion, as such experiments take into account the essential load-bearing 
role of skeletal tissues, they should become a key aspect of any investigation 
into bone cell biology. Thus, it may be considered appropriate that studies into 
the effects on nonmechanical factors should ideally also be conducted in an 
environment in which skeletal tissues (or indeed other load-bearing connective 
tissues), and cells derived from these tissues experience their normal physi- 
ological range of mechanical stimuli. 

We first address some of the aspects of in vitro culture that need to be con- 
sidered during the interpretation of results derived from experiments investi- 
gating the cellular responses to mechanical stimuli, with respect to their in 
vivo relevance. We then describe model cell culture systems for investigating 
the response of isolated bone cells to mechanical strain, and finally we describe 
methods for loading bone explants in culture. 

1. 1. Limitations of In Vitro Organ and In Vitro Cell Culture Strain 
Application Models 

-m- Organ culture, the maintenance of tissue explants in vitro, is an attempt to 

bridge the gap between cell culture and in vivo models. Many tissues including 
cartilage, tendon, and bone have been studied in organ culture (2-6). This 
offers, as a major advantage, the maintenance of an intact extracellular matrix 
(ECM). ECMs are the product of resident cells, are structurally unique, and are 
specialized according to local requirements. For example, mineralization is 
important for bone rigidity and type I collagen endows it with mechanical 
strength (7,8). Retention of the ECM is important as it maintains normal cell 
attachment sites and spatial relationship between cells in a tissue. By retaining 
the tissue's architectural organization, it is also likely that the relationship 
between distinct mechanical sequelae of loading, such as strain, fluid shear 
stress, and streaming potentials, will be conserved. Both organ and monolayer 
cell cultures allow the responses to be investigated without the complication of 
systemic factors. Yet at the same time it is possible, indeed probable, that some 
of these responses to mechanical stimulus will be complicated by the conse- 
quence of tissue or cell isolation and maintenance in culture. 

Connective tissues, including bone, play major structural roles; for example, 
the skeleton must support body weight, facilitate movement, and protect inter- 
nal organs. In such tissues, the ECM and not the resident cells, fulfils such 
roles; thus its preservation in culture may be considered appropriate in experi- 
ments that are aimed at establishing the basis of the tissue's response to load- 
bearing. Bones undergo an adaptive response to dynamic loading in vivo; with 

29/Pitsillides/399-422/F1 400 1 2/26/03, 10:56 AM 

I ■ I I I 

Bone Cell Loading in Culture 401 

very low cycle numbers capable of inducing changes in bone architecture and 
mass (9-11). Indeed, an osteogenic/antiresorptive regimen of 36 loading cycles 
at 0.5 Hz is sufficient to produce a maximal response (9). The brief duration 
(72 sec) of this mechanical stimulation, required to activate a full osteogenic 
response in vivo, makes the study of the preceding "preosteogenic" events fea- 
sible. Thus, similar loading regimens have been used in organ culture and in 
"strain models" in cell culture to investigate loading-induced responses 
(5,6,12). Because strain, fluid shear stress, and streaming potentials have all 
been reported to influence bone cell metabolism, it is important to appreciate 
that the ECM may modify both the signals to which resident cells respond, as 
well as the specific cellular reaction that such signals generate (9-11,13-15). 

Many bone ECM molecules contain integrin-binding Arg-Gly-Asp (RGD) 
sequences and in vivo changes in integrin expression profoundly affect bone 
metabolism (16-19). This is important, as any lack of integrin-binding sites in 
cell monolayer culture may also lead to changes in integrin expression. ECM 
molecules are known to influence cell behavior. Indeed, osteoblasts that bind 
preferentially to fibronectin in vitro, also exhibit a faster rate of proliferation 
on this substrate than when seeded onto poly-L-lysine (20,21). Differential 
integrin expression has been shown in osteoclasts seeded on bone sialoprotein- 
-®- coated glass compared to plastic (22). It is therefore pertinent, at least in the 

short term, that organ culture maintains normal cell-ECM attachment sites, 
retains cell-cell associations and their three-dimensional relationships. In con- 
trast, cell cultures are usually two-dimensional monolayers in which there is 
little control of such relationships. 

A feature of bone explant culture is that it retains the relative positions of 
osteocytes, osteoblasts, and osteoclasts. These might be important for main- 
taining signaling gradients, for example, all bone cells produce nitric oxide and 
its release may establish local concentration gradients that regulate behavior 
(23). It is unlikely that such gradients are produced in cell culture models. 
Further, osteocytes and osteoblasts can communicate via gap junctions found 
at the termini of cell processes stretching through the canaliculi (24). Despite 
the fact that these are observed in cell culture (25), any positional information 
that they may confer will be lost as their expression patterns are altered as a 
consequence of their two-dimensional organization. These arguments may be 
particularly pertinent for the terminally differentiated osteocytes, which are 
normally embedded in the bone ECM and are difficult to isolate and grow in 
cell culture. Although an osteocyte-like cell line has recently been produced, at 
present there is only one well-established method for isolating primary osteo- 
cytes for cell culture (26) (see the chapter by Nijweide et al., this volume). It 
may also be relevant that as well as maintaining the relative position of bone 
cell types, organ culture models also conserve their normal cell ratios. Thus, 

29/Pitsillides/399-422/F1 401 1 2/26/03, 10:56 AM 

I ■ I I I 

402 Pitsillides et al. 

organ culture models can be used to examine potential cellular crosstalk, which 
may occur during the tissue's response to mechanical stimuli. In contrast, cell 
culture can only partially achieve this through the use of coculture systems in 
which cell ratios and positional relationships will only approximate to those in 

1.2. The Effects of Loading: Mechanical Strain, Fluid Shear, 
and Streaming Potentials 

Retention of structural integrity in organ culture allows the response to 
applied loads that are capable of generating physiological levels of mechanical 
strain to be investigated. Ideally, cell culture would complement this by allow- 
ing for particular responses to a specific single component, uniaxial strains, 
applied to large numbers of uniform cells, to be determined. However, logisti- 
cal problems and the fact that substrates stretched along one axis will inevita- 
bly contract at 90° to this principle strain axis (subject to Poisson's ratio) mean 
that such "ideals" are not readily achievable. As this ratio is fixed for any single 
material, but can vary between different materials, direct comparison between 
studies in which substrates differ should be made only with care. Currently, 
such considerations are unavoidable. With appropriate loading, cells in explants 
-0- can probably experience physiological levels of strain at their given location. 

However, despite precise loading regimens, cells within explants will experi- 
ence a range of strain magnitudes in either compression or tension and these 
location-specific variations should ideally be taken into account during experi- 
mental design and interpretation. 

Another consideration involves the fact that the lacuna-canalicular network 
of bone is filled with tissue fluid, and that although this network is retained in 
organ culture models, its contents are replaced by culture media. When this 
fluid flows over, or through the ECM, fluid shear forces are generated. Appli- 
cation of shear forces in vitro has been shown to induce a response in endothe- 
lial cells, chondrocytes, and bone cells (13,14,27,28). In bone, fluid shear forces 
are generated as a result of a mechanical load-induced shift in tissue fluid, and 
thus mechanical strain cannot be investigated independently of fluid shear in 
vivo or in organ culture models. It is possible to investigate the independent 
effects of fluid shear application in cell culture systems (see the chapter by 
Stevens and Frangos et al., this volume). Nevertheless, it is also clear that the 
ECM modifies fluid shear rate and magnitude in a way that it is practically 
impossible to reproduce currently in cell culture models. 

A further complication is introduced by the contribution of the flow of 
charged tissue fluid over the surface of bone's charged ECM, which results in 
electrical, streaming potential, currents. Media bathing lacuna-canalicular net- 
works in bone explants contain many charged molecules, including amino acids 

29/Pitsillides/399-422/F1 402 I 2/26/03, 10:56 AM 

I ■ I I I 

Bone Cell Loading in Culture 403 

and proteins. Thus, both the media and the surface of the tissue show a net 
charge, and any potential difference created at this site of contact results in an 
electrostatically charged layer around the tissue. In intact bone, cyclical 
mechanical loading creates fluid flow and establishes a streaming potential. 
Removal of bone from an animal isolates it from the load it normally experi- 
ences and results in fluid shifts. It is therefore important to acknowledge the 
disturbance to these streaming potentials that are an unavoidable consequence 
in both organ and monolayer cultures. 

Cell cultures have as their main advantage that they are universally estab- 
lished, easy, and reproducible (see Parts I and II). 

The media, cell number, percentage confluence and differentiation state 
can all be defined and controlled, making it possible to reproduce experi- 
ments in different laboratories. Similar standardization is not yet achieved in 
organ cultures, but should be achievable in the future by use of the same 
culture media, explants of similar size, origin, age, sex, hormonal status, and 
so forth. In contrast with cell cultures, it is almost impossible to control cell 
content, tissue architecture or sample heterogeneity in organ cultures, and 
the responses within any one cell type is likely to be subject to paracrine 
controlling influences. Although recent developments have been made (see 
-®- Subheading 3.5.), another current drawback of organ culture models is the 

limited time that explants can retain viability ex vivo. Finally, loadable organ 
cultures are clearly not appropriate for the application of all techniques. For 
instance, mechanically stimulated increases in intracellular free calcium are 
impossible to monitor in organ culture, but can be relatively easily observed 
in cell monolayer culture. 

2. In Vitro Cell Culture Loading Models 

Before describing any specific methods (see Note 1), we emphasize that the 
history of the cell-straining techniques has, for the most part, relied on custom- 
built devices that remained relatively unique to individual investigators. To 
some extent this may have been due to a desire for progressive augmentation in 
the system's efficiency in delivering a specific component of the mechanical 
loading environment. This is clearly a desirable objective. Nonetheless, it does 
not negate the contribution to our understanding made by systems in which the 
precise nature of the stimulus to which cells are exposed remains largely unde- 
fined. It is, however, surprising, in our opinion, that few if any of these studies 
have attempted to relate zonal variations in the applied stimulus to any differ- 
ences in the cell's biochemical response at the level of an individual cell. To do 
this, the investigator would simply need to measure some response, at the level 
of an individual cell, and relate this to the magnitude of the mechanical stimu- 
lus to which that cell was exposed. 

29/Pitsillides/399-422/F1 403 1 2/26/03, 10:56 AM 

I ■ I I I 


404 Pitsillides et al. 

In an attempt to delineate the defining character of straining techniques, it is 
important that a number of principles are considered: 

1. A mechanical strain (often tensional) is applied by substrate deformation. 

2. A known strain magnitude is generated in the substrate onto which the cells are 

3. The substrate is able to support cell growth and differentiation. 

4. Cells must attach to the substrate and not detach as a consequence of defor- 

5. The substrate should have perfect elastic properties. 

6. Adequate access for observation, extraction, and other processing is possible. 

7. During experiments, the apparatus should be housed in a controlled gaseous and 
thermostatic environment. 

8. Techniques should aim to subject all cells to equal levels of mechanical strain, or, 
as a minimum, ensure that their levels are accurately measured or estimated and 

The extrinsic mechanical stimulus to which bones respond may be a change 
in tension, compression, gravitation, vibration, or hydrostatic pressure. Herein, 
we outline different techniques that have been developed to study the response 
of cultured cells to tension in culture, by stretching (or bending) the substrate 
onto which the cells have adhered. Thereafter, we will describe in detail one 
such four-point straining device that we have used extensively. 

2. 1. Biaxial Straining 

In "dish-deforming" cell culture models, the imposed biaxial substrate 
deformation provides a nonhomogeneous strain on cells. Commonly, dishes 
are intermittently deformed over a template to produce a 5% change in sur- 
face area (29,30), such that cells near the center of the dish would be 
strained in excess of those at the periphery. Several systems have been 
developed (single- and multiple-dish versions) that control the input strains 
by varying the curvature of the template. In these systems, estimates of the 
template's arc length of spherical distension provide a basis for calculating 
average strains (31). 

Numerous other platen-driven devices for applying biaxial strain have been 
developed for cells grown in culture dishes with a flexible membrane, instead 
of rigid tissue culture plastic, bottoms. Examples of such techniques include: 
upward or downward "tenting" of membranes using vertically pulsating ball- 
ended prongs (32,33), electrocy finder-driven pulsations of spherical watchglass 
sectors to indent culture dishes (34), and upward indentation of Petri dishes 
with flat-ended circular pistons (35). 

In 1985, flexible-bottomed circular, cell culture plates were specifically 
designed that could be interfaced with a computer-controlled vacuum manifold 
system (36). Vacuum application to the culture well undersurface stretched it 

29/Pitsillides/399-422/F1 404 I 2/26/03, 10:56 AM 

I ■ I I I 

Bone Cell Loading in Culture 405 

downward in a manner that could be controlled by varying vacuum magnitude, 
waveform, frequency, and number. Another differential pressure, flexible-sub- 
strate system used positive, solenoid valve-mediated, pressure to deform circu- 
lar, peripherally clamped, lOO-^m polyurethane-urea membranes (37). 

A common problem with such devices is that the deformation can result in 
nonuniform strains of the substrate. It is clear that these will also be subject to 
variations in the strain applied to cells at various positions across their diam- 
eter. This is an important consideration, as it has been predicted that biaxial 
straining is at least twice as potent as uniaxial strains (38). Thus, unless one 
was going to investigate a particular response in a single cell of known location 
and strain, then it would be necessary to appreciate that any results may repre- 
sent an aggregate of total inhomogeneous cellular responses to the applied 
range of strains. One might also consider the effect of "puddling"; that is, as 
the substrate is drawn downwards by the vacuum, the overlying medium will 
collect at the lowest point. The effect of this fluid movement has not been 

Another method that aims to apply uniform biaxial strain involves using a 
square elastic substrate that is stretched on four sides, over the rim of a platen. 
As increasing amounts of the original (flat) membrane slip radially, outward 
-W~ over an axially advancing platen rim (its area increasing as it does), the portion 

of the membrane remaining directly over the platen theoretically experiences 
homogeneous biaxial strain. Brighton et al. found that when straining cultured 
osteoblasts in this way, they could detect a physiological response at strains as 
low as 300 fx£ (39). 

2.2. Uniaxial Straining 

These experiments purport to investigate the effect of uniaxial strains, how- 
ever, it is important to note that they nonetheless are subject to the restrictions 
imposed by Poisson's ratio (a substrate stretched along one axis will inevitably 
contract at 90° to this principle strain). Most of these devices have relied on 
four-point bending, but a three-point bending technique was designed by 
Hasegawa et al. (30). Whilst this successfully applied uniaxial strains, their 
magnitude over the plate's surface will have been unequal. 

2.2.1. Stretching of Substrates 

Uniaxial straining devices have been developed in which a strip of silicone 
(see Note 2) is placed in a trough and anchored at one end, a magnet attached at 
the strip's free end and another magnet outside the culture system, operated by 
hand or via a motor-driven cam, is used to stretch the substrate. Other methods 
using silicone as a substrate require film spools to create the length change 
(38). Using derivatives of such systems, in which cells are seeded onto 

29/Pitsillides/399-422/F1 405 1 2/26/03, 10:56 AM 

I ■ I I I 

406 Pitsillides et al. 

prestretched membranes, it is possible to observe cells during their response to 
compression as well as tension. 

Recently, cell culture chambers attached with silicone sealant to the surface 
of polycarbonate sheets have been used to subject adherent cells to mechanical 
strain (40). This system, based on a design originally described by Murray and 
Rushton (41), uses strips of polycarbonate that are adapted into cell culture 
chambers, by attaching lids removed from four-well slides with silicone 
sealant. Cell straining is achieved using a device consisting of two platens 
that run on dry linear bearings, driven by a servo-controlled pneumatic ram. 
The actuator and compressor are housed outside the incubator and the 
former is controlled from a computer supplied with feedback from a linear 
variable displacement transducer. The polycarbonate strips, with adherent 
wells and cells, are clamped across the two platens. Controls consist of 
strips clamped across "static" platens as well as strips that are clamped 
across a moving platen that will generate medium perturbation equivalent 
to that experienced by strips subjected to strain stimuli (40). This device is 
capable (determined by appropriate strain gauge measurements) of apply- 
ing controlled cyclical strains between 100 and 200,000 [xe, at strain rates 
from 100 to 1,000,000 fxe/sec and in this system, strains have been applied 
-®- as a ramped square wave pattern. 

Similar systems for applying uniaxial strain using a polyurethane substrate 
have also been developed by Grabner et al. (42), in which a motor-driven lin- 
ear stage was used for the application of a cyclical tension. These devices 
appear to represent a key to addressing questions that relate to the differential 
effects of mechanical strain and fluid flow. 

2.2.2 Four-Point Bending of Substrates 

Four-point bending of a plate requires a pair of lateral forces acting out- 
side two fulcra. This produces an even curve at both the tension (convex) 
and compression (concave) surfaces, and the curve makes up a segment of 
a circle, between the fulcra. Several four-point bending systems have been 
designed, and in the simplest terms these differ in the manner by which 
loads are applied, and the culture substrates compounds that are used. One 
such system, which uses rectangular culture plates (polycarbonate or glass), 
was recently devised in which culture plates are suspended within a com- 
mon 40 x 40 mm silicone rubber well (43). By allowing both sides of the 
substrate to act as the surface onto which cells can be seeded, the device 
elegantly permits the cell's response to either tensile or compressive strains 
to be examined. It also includes scope for direct strain gauge readout to 
monitor the parameters of strain to which cells are exposed. Another pow- 
erful feature of this device is that it permits the use of a range of culture 

29/Pitsillides/399-422/F1 406 1 2/26/03, 10:56 AM 

I ■ I I I 

Bone Cell Loading in Culture 407 

substrate materials, with different Young's moduli, thus facilitating appli- 
cation of a broad range of strain magnitudes. 

2.3. An Example of Four-Point Bending 
of Monolayers of Cells on Plastic Strips 

A four-point bending system that is able to engender low-strain levels has 
been used (5,44-48), in which tensional strain is applied to plastic strips con- 
taining adherent monolayer bone cell cultures in a custom-designed loading 
apparatus (Fig. 1). This four-point bending system was designed to deliver 
strain levels of several hundred to several thousand ^ie, which are within the 
physiological range recorded for bone cells in vivo (see Note 3). It should be 
noted that this system is also capable of delivering compressive strain. The 
method is as follows: 

1. Passage cells onto presterilized plastic strips for at least 24 h (this time depends 
upon initial seeding density) and maintain cells for 24 h and throughout the 
experimental application of mechanical strain in a humidified atmosphere of 95% 
air-5% C0 2 in serum-free medium (see Note 2). 

2. Transfer the strips (n = 5 for each variable) to the loading apparatus under sterile 
conditions. To do this, separate the loading apparatus into its two component 

-{©)- parts: a base for the strained strips and an upper portion with the cam, platen, and 

flow chambers. Remove strips from the dishes in which they have been 
preincubated, and position into individual chambers with the requisite volume of 
medium (10 mL) added. 

3. Reconstruct the whole apparatus, place back into the humidified incubator, and 
allow for equilibration. Keep disturbance of the medium minimal at all times. 

4. Apply a load generating mechanical strain of 3400 u,e at 1 Hz for 600 cycles to 
cells adherent on the strips (see Note 3). 

5. Subject similar strips with cells attached to cyclic perturbation of the medium 
(flow controls) without applied loads, and subject others only to identical changes 
of the medium without any mechanical perturbation to serve as "static" controls 
(see Note 4). 

6. Following strain application, remove strips from loading apparatus and return to 
dishes for various times post-straining, depending on response to be investigated. 

2.4. Alternative Methods for Applying Strain 

Magnets are used to provide a high degree of control in devices capable of 
stretching a substrate, and such magnetostrictive actuators have been used in 
order to generate up to 22,000 \i£ on particular substrates. The electromagnetic 
elements placed at either end of the substrate are attached to clamps mounted on 
a guide plate that ensures unidirectional travel. It is an aspect of these devices 
that they generate powerful electromagnetic fields that may directly affect cell 
behavior, and shielding from these fields should always be used. 

29/Pitsillides/399-422/F1 407 I 2/26/03, 10:56 AM 

I ■ I I I 


Pitsillides et al. 








Flow (control) 






Fig. 1. Schematic diagram of the custom-built apparatus used for straining cells by 
subjecting plastic strips to four-point bending. As the eccentric cam rotates it presses 
the loading platen onto the vertical edges of the plastic strips and causes their defor- 
mation (to the dotted position) in an arc around the paired fulcra underlying the strips 
(white). The platen is returned to the unloaded position by the spring. The rotation of 
the cam also acts to rock the chamber into which strips to be subjected to flow "con- 
trol" are placed. For dose-response experiments, peak strain magnitudes can be al- 
tered by vertical displacement of the base unit by using a range of washers (with 
different thickness, placed between base and upper units; see Fig. 4), and frequency 
altered with rheostat control of cam revolution rate. This unfortunately affects the 
waveform produced. However, a novel device has recently been designed (yet un- 
tried), in which the platen's vertical displacement is governed by a screw mechanism 
that is computer controlled. This offers greater flexibility in the waveforms that can be 
generated, such that on/off dwell times and the "on" as well as "off strain rates can be 
varied independently (see Fig. 5; Stromberg et al., personal communication). In addi- 
tion this model utilizes commercially available cell culture strips (of equivalent size to 
microscope slides) that do not possess the end walls. 

As an alternative, piezoelectric extension may be used to provide the dis- 
placement. This allows accurate control, however, like magnetostrictive actua- 
tors, generates high electrical fields. Such actuators have been used in vitro to 
apply strains of controlled magnitude (200^-0,000 \i£), frequency (up to 100 Hz), 



2/26/03, 10:56 AM 

I ■ I I I 

Bone Cell Loading in Culture 409 

and waveform, which adequately cover the range experienced by bone cells in 
vivo. The cells are mechanically strained by moving a plunger, connected to 
the center of an actuator, both ends of which are inserted into grooves on an 
acrylic resin frame, thus generating maximum actuator displacement. Cultured 
cells are seeded on, or in, a collagen gel block between plungers made of a non- 
conductive acrylic resin material, and the gel block anchored by stainless-steel 
wire meshes (lattice size; 0.4 x 0.4 mm) fixed to the plunger-ends (49). 

3. In Vitro Organ Culture Loading Models 

Culturing bone as a tissue allows for the preservation of normal cell-matrix 
attachments as well as cell-cell attachments between resident cells. Preserva- 
tion of the structural, load-bearing, mineralized compartment means that load- 
ing of segments in vitro can be controlled so that the extent of bending 
produced, engenders levels of strain (measured directly in loaded segments 
using attached strain gauges) identical to those measured at the same site dur- 
ing normal physiological activity in vivo (see Note 3). Providing disruption is 
kept to a minimum, it is also likely that in vitro loading of bone segments will 
reproduce other associated phenomena of loading (fluid shear/streaming 
potentials). On the other hand, such isolation will result in many unavoidable 
-W~ changes, including loss of blood flow, a potential restriction on the ability of 

nutrients to reach cells, and changes in the relationship between bone and its 
marrow. Nonetheless, this approach allows attempts at bridging the gap 
between the morphological disciplines and those of the cell biologist, in deter- 
mining the mechanisms of bone's response to loading (see Note 4). 

The in vivo loadable avian ulna model developed by Lanyon and Rubin 
(9,50) determined that dynamic mechanical loads applied to a functionally iso- 
lated portion of bone could result in an increase in bone formation. The extent 
of new bone produced, was dependent on the magnitude and the rate at which 
the strains were engendered (51). Further studies indicated that only 36 cycles 
(over a period of 72 sec) per day of applied load were sufficient to produce a 
maximal formative response (9). That all the information needed to produce 
new bone was contained in only a short period of time, provided the rationale 
for generating organ culture models to investigate other early loading-related 
responses that might contribute to controlling the osteogenic response. 

The in vitro loadable models, which have been developed in Lanyon's labo- 
ratory, include adult canine cancellous bone cores, rat cortical ulnar bones, rat 
parietal bones, and embryonic chick tibiotarsi. All of these have used the afore- 
mentioned in vivo studies and have, as a consequence, used similar strain regi- 
mens as the mechanical stimulus. Further models from other groups include a 
loadable bovine cancellous bone core system (52) and a human bone core model 
(53). Each system requires that attendant soft tissues and marrow (where pos- 

29/Pitsillides/399-422/F1 409 I 2/26/03, 10:56 AM 

I ■ I I I 


410 Pitsi Hides et al. 

sible) are removed, to leave exclusively bone tissue with resident bone cells and 
associated internal vasculature. To generate results that are likely to have in vivo 
relevance, it is imperative that these organ culture systems are calibrated. Thus, 
prior to loading bone in vitro, the amount of force required to generate those 
levels of strain that are engendered by normal activity in vivo, should be defined 
(see Notes 3, 5, and 6). In the following subheading we describe several models 
of loadable bone organ cultures that have been defined in this way. 

3. 1. Loading of Adult Canine Cancellous Bone Cores 

3.1.1. Reagents 

1. Culture medium: Minimum essential medium (MEM) + Hanks' salts and 25 mM 
Af-2-hydroxyethylpiperazine-W-2-ethanesulfonic acid (HEPES) (Gibco) supple- 
mented with 2.0 mM L-glutamine (Gibco), 0.1% bovine serum albumin (Sigma), 
100 IU/mL of penicillin, and 100 ng/mL of streptomycin (Gibco). Perform all 
cultures at 37°C in an air incubator. 

3.1.2. Preparation of Bone Cores 

1 . Euthanize the animal with an overdose of barbiturate, flex the stifle, and make an 
incision longitudinally in the skin. 

2. Deflect the kneecap to expose the trochlear groove of the distal femur. Using a 
trephine with a nonretractable pointed insert (slightly proud of the cutting edge), 
make a circular groove in the articular cartilage surface. Remove the pointed 
insert and take a full-depth cartilage and bone plug from the epiphysis of both left 
and right limbs. 

3. Place each core into a cutting rig, and trim to a uniform length (1 cm). 

4. Introduce the core into a syringe barrel with the nozzle portion and end removed. 
Flush sterile, warm phosphate-buffered saline (PBS) through the core to remove 
marrow from between the bony trabeculae. 

5. Store cores individually in culture medium in sterile bottles at 37°C until all cores 
required for an experiment have been collected. 

3.1.3. Loading of Bone Cores 

1. Within separate flexible collars of silicone tubing (see Note 7), support each of 
the bone cores between a pair of milled Perspex supports, which have been drilled 
to allow the passage of culture medium through them and the core (see Note 5). 
Make sure the end face of each support is cut to be of equal diameter to the core. 
To create a seal between the supports and to surround the core, the internal bore 
of the silicone tubing should be equal to the core diameter. Hold the control chan- 
nels in place on a backboard with clips. 

2. Place cores into the loading apparatus (Fig. 2). To do this, mount the lower 
Perspex support onto a rigid, nonflexible, bar attached to the backboard. Connect 
the upper support to a pneumatically operated actuator, fixed to the backboard, 

29/Pitsillides/399-422/F1 410 1 2/26/03, 10:56 AM 

I ■ I I I 

Bone Cell Loading in Culture 





Fig. 2. Loading apparatus for bone cores. 

which is operated by an air compressor unit. Introduce recirculating medium into 
the reservoir syringe above each core and draw through the silicone tubing, by a 
peristaltic pump, back to the reservoir syringe above the core (one lane per core). 
Controls comprise cores housed similarly in silicone tubing and Perspex sup- 
ports, which are attached to the backboard with clips but remain unloaded. 
Prior to loading, preincubate all cores for 4-5 h in a recirculating system, in 
which culture medium is delivered at a flow rate of 0.3 mL/min. Following this 
preincubation period, convert the recirculating system to single-passage perfu- 
sate mode. Refill all reservoirs and replenish continually for the remainder of the 
experiment for any "real time" analyses. Alternatively, retain the recirculating 
system as appropriate. 

To load bone cores, regulate air pressure to deliver the same force that generates 
a bulk strain of 5000 [xe in separate cores, at a loading frequency of 1 Hz. 

3.2. Rat Ulnae Organ Cultures (see Note 6) 

3.2.1. Reagents 

1. Culture medium: Dulbecco's modified Eagle medium (DMEM) plus 10% char- 
coal-dextran extracted fetal calf serum, supplemented with 2.0 mM L-glutamine 
(Gibco), 100 IU/mL of penicillin, and 100 ng/mL of streptomycin (Gibco). Per- 
form all cultures at 37°C in a humidified incubator in 5% C0 2 at 37°C. The 
culture medium used in these experiments depends on the conditions employed 
during the application of loading (see Subheading 3.2.3.). 


2/26/03, 10:56 AM 

I ■ I I I 


412 Pitsi Hides et al. 

3.2.2. Preparation of Bones for Loading 

1. Dissect ulnae and clear of attendant soft tissue. Cut ulnae to equal length and 
remove marrow. 

2. Preincubate on hand-milled polytetrafluoroethylene (PTFE) supports at the air- 
medium interface for 6 h in 12-well culture plates in a humidified atmosphere of 
95% air-5% C0 2 at 37°C. 

3. Place into one of two different loading devices (weight lifting, or pneumatic ac- 
tuator; see Subheading 3.2.3. and 3.2.4.). Following loading, remove bones from 
the loading apparatus and return to PTFE supports in 12-well plates for various 
times post-loading, depending on the response to be investigated. 

3.2.3. Loading of Bones in Weight-Lifting Model 

1. Ulnae are loaded in individual chambers with medium recirculating around the 
cortical shaft. This loading apparatus is maintained in an air incubator, so during 
the loading period, the culture medium is supplemented with 25 n\M HEPES buffer. 

2. Each bone is held vertically between two cups. The lower cup is connected to an 
eccentric cam, while the upper cup is connected to a weight-carrying platform 
that rests on a ledge. As the cam rotates, the bone is raised and lowered. On the up 
stroke, the bone acts as the sole support for the weight-carrying platform and on 
the down stroke the weight returns to rest upon the ledge. The amount of weight 
on the platform can be altered, thus permitting different levels of mechanical 
strain to be engendered (12,54,55). 

3. Following loading, remove bones from the loading apparatus and return to 12-well 
plates for various times post-loading, depending on the response to be investigated. 
Controls comprise ulnar segments, treated identically, that do not lift weights. 

3.2.4. Loading of Bones in the Pneumatic Model 

1. This rat ulna organ culture model maintains the tissue in a fixed volume of me- 
dium (see Fig. 3). The device consists of a milled polycarbonate block contain- 
ing 10 chambers, in which five are customized to permit loading and five serve as 
controls; loads are applied by pneumatic actuators. 

2. The bone shafts are removed from the preincubation medium and introduced into 
the holding cups of the loading apparatus, either in loading or control chambers 
each containing 4 mL of culture medium. 

3. After 5 min equilibration time, bone explants are loaded axially to generate 
mechanical strains levels on the lateral mid-shaft, similar to those generated in 
vivo. Control bone shafts are not loaded. 

4. After the loading period, bone shafts are transferred back to 12-well plates and 
cultured for various periods of time, depending on the response to be investigated 

3.3. Loadable Rat Calvariae Organ Cultures 

3.3.1. Materials 

The culture medium is as given in Subheading 3.2. 

29/Pitsillides/399-422/F1 412 I 2/26/03, 10:56 AM 

I ■ I I I 

Bone Cell Loading in Culture 










Fig. 3. Diagram of organ culture loading apparatus, showing the bone segment held 
between nylon caps immersed in culture medium. Oscillating air pressure in the pneu- 
matically operated cylinders applies a dynamic load to the tissues. 


3.3.2. Method 

Essentially the method used is identical to that described under Subheading 

3.2., but an adaptation is made in the caps that hold the tissue within the appa- 
ratus (6). The caps retain bones to limit uncontrolled translation. For the ulnar 
bone shafts these nylon caps are fashioned into cups, whereas for the parietal 
bone (calvariae) explants, the caps have a shallow, narrow rebate engineered to 
accommodate the thin, plate-like architecture of these bones. 

1. Cut the parietal bones into rectangular explants and culture as described in Sub- 
heading 3.2. 

2. To apply loads, first locate the rostral and caudal ends of the bones into the "cap" 
rebates, thus allowing axial loading without scope for slip. 

3. Perform loading using the pneumatic actuator device (see Subheading 3.2. and 
Notes 6 and 8). 

4. After the loading period, bone expalnts are transferred back to 12- well plates and 
cultured for various periods of time, depending on the response to be investigated. 

3.4. Loading of the Chick Tibio-Tarsus (see Note 9) 

3.4.1. Materials 

1 . Culture medium: Fitton-Jackson's modification of BGJb medium containing 2 mM 
L-glutamine, 100 IU/mL of penicillin, 100 ng/mL of streptomycin, 50 ug/mL of l- 
ascorbic acid, and 2% heat-inactivated fetal calf serum. Culture bone at 37°C in a 
humidified 5% C0 2 incubator (see Subheading 3.2.). 



2/26/03, 10:56 AM 

I ■ I I I 


414 Pitsi Hides et al. 

3.4.2. Method Preparation of Bone Segments 

1. Euthanize 18-d-old embryonic White Leghorn chicks using an approved method 
and remove the tibiotarsi. 

2. Remove adherent soft tissues and fibula, and both cartilaginous ends to leave a 9-mm 
bone shaft segment. 

3. Aspirate the marrow, but leave the periosteum intact. 

4. Wash briefly in PBS. 

5. Culture bone segments at the air-medium interface on PTFE supports for 5 h. Loading of Bone Segments 

1. Hold the embryonic tibial bone shafts at either end by polypropylene caps, in 
chambers milled into a Perspex block. 

2. Both "weight lifting" and "pneumatic" actuator devices (58) have been used to 
load these bones (Subheading 3.2.). Each device is calibrated to generate the 
required strain levels (see Note 2). 

3. For weight-bearing induced loading, an eccentric cam (rotating at 1 Hz) is 
employed to raise and lower an L-shaped cradle holding the weights. The force is 
transferred to the bones via a pivot positioned at the right angle of the L-shaped 
cradle, and thus when the cradle is lowered, the tibia supports the weight of the 
cradle (59,60). As in other models, the level of engendered strain can be altered 
by changing the amount of weight that is "lifted." 

3.5. Long-Term Perfusion Loading Model 

A recently described long-term model system designed to overcome the lim- 
ited viability of previous organ explant models (53) may soon make significant 
advances in our understanding of loading-related bone formation. This system 
involves perfusion of trabecular bone cores, includes the marrow, and extends 
the lifetime of the tissue to 72 d in vitro (see ref. 61). Preliminary studies using 
the incorporation of fluorescent labels into newly mineralized surfaces, to pro- 
vide a direct measurement of bone formation rate, have shown that these bone 
cores can retain their in vivo rates of bone formation for up to 20 d in vitro 
(52). Along with the loadability of such explants in vitro, it is possible that 
such models will provide a means by which the cellular basis for "sensing" 
mechanical stimuli or "communicating" their influence to coordinate loading- 
induced changes in bone remodeling can be elucidated. 

4. Notes 

1. We highlight in Subheading 2. that systems for applying mechanical strain are 
currently, for the most part, unique to individual laboratories. This makes a 
detailed description of a generic device impossible. For this reason, we have given 
a broad account of some of the many, and various, "models" that have been used 

29/Pitsillides/399-422/F1 414 I 2/26/03, 10:56 AM 

I ■ I I I 


Bone Cell Loading in Culture 415 

to apply mechanical strain to isolated bone cells in vitro (see Subheading 2.). 
We have also tried to give some insights into their strengths and weaknesses, the 
materials used and the methods employed, and some of the considerations that 
should be taken into account when deciding which model might be chosen (see 
Subheading 1.). This choice clearly depends on the hypothesis that each investi- 
gator has selected to address, and is one that must take into account which 
mechanical consequence of loading (strain, flow, streaming potential) is being 
examined. Although each of these can be accurately defined, each also exhibits 
many components that can vary. For example, magnitude, frequency, and rate of 
change can be varied during mechanical strain application, and it is vital that 
each of these is considered prior to selecting a "model" in which bone cell 
responses to are to be examined. Perhaps the most obvious and pertinent of these 
deliberations is whether the applied stimulus falls within the physiological range. 
This can only be reflected upon if the variable in question has been measured 
directly in vivo. 

2. A vital consideration for cell culture experiments is that both the substrate and 
culture wells should be made from biocompatible materials. Indeed, it has been 
found that "biocompatibility" of material positively correlates with the amount 
of noncollagenous matrix protein produced by cultured bone cell monolayers 
(62). Medical grade silicone, with its high biocompatibility, is therefore a ratio- 
nal choice of substrate. Aclar-33C (Allied Chemical Co.) is a clear, biocompatible 
material that may be sectioned for electron microscopy, transmits UV light with 
little attenuation or scatter, and is also useful in such devices. 

3. Load-strain relationships have to be determined in any new model of in vitro 
strain application by prior calibration. Thus, for example, plastic strips identical 
to those that will be used experimentally must previously have had strain gauges 
attached to them and loaded in a controlled manner. During load application to 
these substrates in vitro (and indeed loading of bone segments in organ culture), 
the correct force must be delivered to produce the required level (physiological 
or if required nonphysiological) of mechanical strain. Strain gauges cannot be 
attached to samples that will be used for experimental investigation; therefore it 
is necessary before each new device is used to establish the relationship between 
applied load, which can be monitored and adjusted directly, and the resultant 
strain, which cannot. To calculate the relationship between strain and load for 
plastic strips (and bone explants; see Figs. 4 and 5, [11]) use single-element 
microminiature strain gauges. Bond gauges to the test substance (or bone sur- 
face) with cyanoacrylate adhesive, ensuring that that the prewired strain gauge 
will detect strains along the substrate's principal strain axis. Connect gauges to a 
strain gauge conditioner and amplifier, in a quarter bridge configuration, and 
record the output voltage on a personal computer equipped with an analog to 
digital (A/D board) converter. Convert to strain values using a 1000 \it calibra- 
tion shunt resistance built into the conditioner units. Use the recorded strain data 
files to assess peak strain magnitude and strain rate and determine frequency of a 
loading cycle. 

29/Pitsillides/399-422/F1 415 1 2/26/03, 10:56 AM 

I ■ I I I 


Pitsillides et al. 






0.6 0.8 1 

Washer Thickness (mm) 

Fig. 4. Peak strain magnitude in plastic strips can be altered by changing the differ- 
ence in height between the lower and upper base units in the four-point loading device. 

Fig. 5. Waveform of the loading generated strains in two strain gauges attached to 
one strip in the loading apparatus (Stromberg et al., personal communication). 

4. Mechanical strain application results in the exposure of cells to both tensional 
strains, via bending of their substrate, and perturbation of their medium as a con- 
sequence of the cyclical displacement of these strips vertically through their 
medium. As a precaution, devices are therefore designed with flow "controls," 
which are included to allow identification of both flow-related and non-flow- 
related responses in the cells that have been exposed to mechanical strain. 
Examination of the cellular responses to such flow stimulation (medium pertur- 
bation) has made it clear that this often includes components of the cell's re- 
sponse to strain application, but that it also often induces a range of responses 
that are particular and specific to the application of flow itself (45). Thus, it is 
vital that controls are included that have been subjected to similar "preparatory" 
changes in the medium, but that are not subjected to any form of further medium 
perturbation or substrate deformation (i.e., "static" controls). An essential caveat, 
however, in the flow "controls" included is that the precise nature of the flow 



2/26/03, 10:56 AM 

I ■ I I I 


Bone Cell Loading in Culture 4 1 7 

stimulus to which the cells are exposed remains, unlike in other studies (see the 
chapter by Stevens and Frangos, this volume), ill defined. Many other cell culture 
variables may also affect the cellular response to strain application. These are 
likely to include: the effects of changing the medium, serum deprivation, cell 
density, growth rate, differentiation status, and so forth. Although most devices 
allow the investigator to sample medium at various times after strain application, 
the investigations are often limited by the fact that large volumes of medium are 
required in such studies and by the fact it is very difficult to make direct micro- 
scopic observations in vitro. 

5. Cell viability should be tested in preliminary experiments by measuring intracel- 
lular lactate dehydrogenase activity and the ability to produce cAMP in response 
to parathyroid hormone. (See chapter by Nobel and Stevens, this volume.) In 
addition, the detection of a difference between loaded and control bones indi- 
cates that the cells present in the cultures are still viable (44,63-65). 

6. In these experiments, medium can be supplemented with exogenous factors, such 
as enzyme inhibitors/activators, and their effects on the response to loading 
investigated. It is also possible using these systems to measure the accumulation 
of soluble metabolites in the medium conditioning each individual culture 
(6,44,56,57). An important strength of such models is that they allow detailed 
investigation of the specific changes associated with individual cells within bone 
tissue. Thus, appropriate treatment (e.g., chilling tissue and preparing cryostat 
sections) of individual bone segments, at various times after loading, allows for 
the analysis of specific components of the strain response to be examined at the 
individual cell level by in situ analyses. 

7. When rigid Perspex is used to surround the core, "barreling" of the core is pre- 
vented and no loading-related responses are measurable. When silicone tubing is 
used instead, 'barreling' of the core is allowed when loads are applied and 
changes in biochemical activity in the loaded compared with nonloaded cores are 

8. In our published studies, we first applied a 100 \xe loading episode. This equates 
to strains that are higher than those determined in vivo, as the loading apparatus 
was not sensitive enough to lower the strains to physiological levels. In later 
experiments, loads that generated 1000 \xz were applied. Although this level of 
strain is approx 30 times greater than physiological, the bone did not fracture, 
reflecting the high safety factor in such bones (6). 

9. This model was initially established to ascertain whether embryonic bone 
responded similarly to adult bone. In some experiments, the periosteum was 
removed to elucidate the contribution from the periosteal cells (58). 


We are grateful to the Biotechnology and Biological Sciences Research 
Council for their generous support (V. D-G.), to GlaxoSmithKline for the 
award of a Special Case Studentship and to The Wellcome Trust for their con- 
tribution to the work done in the laboratories of AAP and Lance Lanyon (S. C. 

29/Pitsillides/399-422/F1 417 I 2/26/03, 10:56 AM 

I ■ I I I 


418 Pitsi Hides et al. 

F. R.). We would also like to thank Dr. Gul Zaman for his constructive and 
critical comments. 


1. Bradbeer, J. N. (1992). Cell biology of bone remodelling, in Recent Advances in 
Endocrinology and Metabolism (Edwards, C. R. E. and Lincoln, D. W., eds.), 
Churchhill Livingstone, pp. 95-113. 

2. Bonassar, L. J., Grodzinsky, A. J., Srinivasan, A., Davila, S. G., and Trippel, S. B. 
(2000) Mechanical and physicochemical regulation of the action of insulin-like 
growth factor-I on articular cartilage. Arch. Biochem. Biophys. 379, 57-63. 

3. Bayliss, M. T., Howat, S., Davidson, C., and Dudhia, J. (2000) The organization of 
aggrecan in human articular cartilage. Evidence for age-related changes in the rate 
of aggregation of newly synthesized molecules. /. Biol. Chem. 275, 6321-6327. 

4. Packer, D. L„ Dombi, G. W., Yu, P. Y., Zidel, P., and Sullivan, W. G. (1994) An 
in vitro model of fibroblast activity and adhesion formation during flexor tendon 
healing. /. Hand Surg. [Am.] 19, 769-776. 

5. Zaman, G„ Pitsillides, A. A., Rawlinson, S. C. F., et al. (1999) Mechanical strain 
stimulates nitric oxide production by rapid activation of endothelial nitric oxide 
synthase in osteocytes. /. Bone Miner. Res. 14, 1123-1131. 

6. Rawlinson, S. C. F., Mosley, J. R., Suswillo, R. F., Pitsillides, A. A., and Lanyon, 
L. E. (1995) Calvarial and limb bone cells in organ and monolayer culture do not 
show the same early responses to dynamic mechanical strain. /. Bone Miner. Res. 
10, 1225-1232. 

7. Currey, J. D. (1979) Mechanical properties of bone tissues with greatly differing 
functions. /. Biomech 12, 313-319. 

8. Riggs, C. M., Lanyon, L. E., and Boyde, A. (1993) Functional associations 
between collagen fibre orientation and locomotor strain direction in cortical bone 
of the equine radius. Anat. Embryol. (Berl.) 187, 231-238. 

9. Rubin, C. T. and Lanyon, L. E. (1984) Regulation of bone formation by applied 
dynamic loads. /. Bone Joint Surg. Am. 66, 397-402. 

10. Turner, C. H., Akhter, M. P., Raab, D. M., Kimmel, D. B., and Recker, R. R. 
(1991) A noninvasive, in vivo model for studying strain adaptive bone modeling. 
Bone 12, 73-79. 

11. Mosley, J. R., March, B. M., Lynch, J., and Lanyon, L. E. (1997) Strain magnitude 
related changes in whole bone architecture in growing rats. Bone 20, 191-198. 

12. Cheng, M. Z., Zaman, G., and Lanyon, L. E. (1994) Estrogen enhances the stimu- 
lation of bone collagen synthesis by loading and exogenous prostacyclin, but not 
prostaglandin E2, in organ cultures of rat ulnae. /. Bone Miner. Res. 9, 805-816. 

13. Reich, K. M., Gay, C. V., and Frangos, J. A. (1990) Fluid shear stress as a mediator of 
osteoblast cyclic adenosine monophosphate production. /. Cell Physiol. 143, 100-104. 

14. Reich, K. M. and Frangos, J. A. (1993) Protein kinase C mediates flow-induced 
prostaglandin E2 production in osteoblasts. Calcif. Tissue Int. 52, 62-66. 

15. MacGinitie, L. A., Wu, D. D., and Cochran, G. V. (1993) Streaming potentials in 
healing, remodeling, and intact cortical bone. /. Bone Miner. Res. 8, 1323-1335. 

29/Pitsillides/399-422/F1 418 1 2/26/03, 10:56 AM 

I ■ I I I 


Bone Cell Loading in Culture 419 

16. Ruoslahti, E. and Pierschbacher, M. D. (1987) New perspectives in cell adhesion: 
RGD and integrins. Science 238, 491-497. 

17. Oldberg, A., Franzen, A., and Heinegard, D. (1986) Cloning and sequence analy- 
sis of rat bone sialoprotein (osteopontin) cDNA reveals an Arg-Gly-Asp cell-bind- 
ing sequence. Proc. Natl. Acad. Sci. USA 83, 8819-8823. 

18. Oldberg, A., Franzen, A., and Heinegard, D. (1988) The primary structure of a 
cell-binding bone sialoprotein. /. Biol. Chem. 263, 19430-19432. 

19. Zimmerman, D., Jin, F., Leboy, P., Hardy, S., and Damsky, C. (2000) Impaired 
bone formation in transgenic mice resulting from altered integrin function in os- 
teoblasts. Dev. Biol. 220, 2-15. 

20. Gronfhos, S., Stewart, K., Graves, S. E., Hay, S., and Simmons, P. J. (1997) Integrin expres- 
sion and function on human osteoblast-like cells. /. Bone Miner. Res. 12, 1 189-1 197. 

21. Cowles, E. A., Brailey, L. L., and Gronowicz, G. A. (2000) Integrin-mediated 
signaling regulates AP-1 transcription factors and proliferation in osteoblasts. /. 
Biomed. Mater. Res. 52, 725-737. 

22. Flores, M. E., Heinegard, D., Reinholt, F. P., and Andersson, G. (1996) Bone 
sialoprotein coated on glass and plastic surfaces is recognized by different beta 3 
integrins. Exp. Cell Res. 227, 40-46. 

23. Collin-Osdoby, P., Nickols, G. A., and Osdoby, P. (1995) Bone cell function, regu- 
lation, and communication: a role for nitric oxide. /. Cell Biochem. 57, 399^-08. 

24. Doty, S. B. (1981) Morphological evidence of gap junctions between bone cells. 
Calcif. Tissue Int. 33, 509-512. 

25. Schiller, P. C, Ippolito, G., Balkan, W., Roos, B. A., and Howard, G. A. (2001) 
Gap-junctional communication is required for the maturation process of osteo- 
blastic cells in culture. Bone 28, 362-369. 

26. Aarden, E. M., Nijweide, P. J., van der Plas, A., et al. (1996) Adhesive properties 
of isolated chick osteocytes in vitro. Bone 18, 305-313. 

27. Garcia-Cardena, G., Fan, R., Shah, V., et al. (1998) Dynamic activation of endot- 
helial nitric oxide synthase by Hsp90. Nature 392, 821-824. 

28. Das, P., Schurman, D. J., and Smith, R. L. (1997) Nitric oxide and G proteins 
mediate the response of bovine articular chondrocytes to fluid-induced shear. /. 
Orthop. Res. 15, 87-93. 

29. Binderman, I., Shimshoni, Z., and Somjen, D. (1984) Biochemical pathways 
involved in the translation of physical stimulus into biological message. Calcif. 
Tissue Int. 36 (Suppl. 10, S82-S85. 

30. Hasegawa, S., Sato, S., Saito, S., Suzuki, Y., and Brunette, D. M. (1985) 
Mechanical stretching increases the number of cultured bone cells synthesizing 
DNA and alters their pattern of protein synthesis. Calcif. Tissue Int. 37, 431-436. 

31. Basdra, E. K., Kohl, A., and Komposch, G. (1996) Mechanical stretching of peri- 
odontal ligament fibroblasts — a study on cytoskeletal involvement. /. Orofac. 
Orthop. 57, 24-30. 

32. Vandenburgh, H. H. (1988) A computerized mechanical cell stimulator for tissue cul- 
ture: effects on skeletal muscle organogenesis. In Vitro Cell Dev. Biol. 24, 609-619. 

33. Soma, S., Matsumoto, S., and Yamamoto, T. (1997) Enhancement by conditioned 

29/Pitsillides/399-422/F1 419 I 2/26/03, 10:56 AM 

I ■ I I I 


420 Pitsillides et al. 

medium of stretched calvarial bone cells of the osteoclast-like cell formation 
induced by parathyroid hormone in mouse bone marrow cultures. Arch. Oral Biol. 

34. Andersen, K. L. and Norton, L. A. (1991) A device for the application of known 
simulated orthodontic forces to human cells in vitro. /. Biomech. 24, 649-654. 

35. Matsuo, T., Uchida, H., and Matsuo, N. (1996) Bovine and porcine trabecular 
cells produce prostaglandin F2 alpha in response to cyclic mechanical stretching. 
Jpn. J. Ophthalmol. 40, 289-296. 

36. Banes, A. J., Gilbert, J., Taylor, D., and Monbureau, O. (1985) A new vacuum- 
operated stress-providing instrument that applies static or variable duration cyclic 
tension or compression to cells in vitro. /. Cell Sci. 75, 35-42. 

37. Winston, F. K., Macarak, E. J., Gorfien, S. F., and Thibault, L. E. (1989) A system 
to reproduce and quantify the biomechanical environment of the cell. /. Appl. 
Physiol. 67, 397-405. 

38. Jones, D. B., Leivseth, G., Sawada, Y., van der Sloten, J., and Bingmann, D. 
(1994). Application of homogenous, Defined strains to cell cultures, in Biome- 
chanics and Cells (Lyall, R. and El-Haj, A. J., eds.), Cambridge University Press, 
Cambridge, UK, pp. 197-219. 

39. Brighton, C. T., Strafford, B., Gross, S. B., Leatherwood, D. F., Williams, J. L., 
and Pollack, S. R. (1991) The proliferative and synthetic response of isolated cal- 
varial bone cells of rats to cyclic biaxial mechanical strain. /. Bone Joint Surg. 
Am. 73, 320-331. 

40. Fermor, B., Gundle, R., Evans, M., Emerton, M., Pocock, A., and Murray, D. 
(1998) Primary human osteoblast proliferation and prostaglandin E2 release in 
response to mechanical strain in vitro. Bone 22, 637-643. 

41. Murray, D. W. and Rushton, N. (1990) The effect of strain on bone cell prostag- 
landin E2 release: a new experimental method. Calcif. Tissue Int. 47, 35-39. 

42. Grabner, B., Varga, F., Glantschnig, H., et al. (1999) A new in vitro system for 
applying uniaxial strain on cell cultures. Calcif. Tissue Int. 64 (Suppl. 1), SI 14. 

43. Jones, D. B., Nolte, H., Scholubbers, J. G., Turner, E., and Veltel, D. (1991) Bio- 
chemical signal transduction of mechanical strain in osteoblast-like cells. 
Biomaterials 12, 101-110. 

44. Pitsillides, A. A., Rawlinson, S. C. F., Suswillo, R. F., Bourrin, S., Zaman, G., and 
Lanyon, L. E. (1995) Mechanical strain-induced NO production by bone cells: a 
possible role in adaptive bone (re)modeling? FASEB J. 9, 1614-1622. 

45. Jessop, H. L., Sjoberg, M., Cheng, M. Z., Zaman, G., Wheeler-Jones, C. P., and 
Lanyon, L. E. (2001) Mechanical strain and estrogen activate estrogen receptor 
alpha in bone cells. /. Bone Miner. Res. 16, 1045-1055. 

46. Zaman, G., Cheng, M. Z., Jessop, H. L., White, R., and Lanyon, L. E. (2000) 
Mechanical strain activates estrogen response elements in bone cells. Bone 27, 

47. Cheng, M. Z., Zaman, G., Rawlinson, S. C. F., Mohan, S., Baylink, D. J., and 
Lanyon, L. E. (1999) Mechanical strain stimulates ROS cell proliferation through 
IGF-II and estrogen through IGF-I. /. Bone Miner. Res. 14, 1742-1750. 

29/Pitsillides/399-422/F1 420 I 2/26/03, 10:56 AM 

I ■ I I I 


Bone Cell Loading in Culture 421 

48. Zaman, G., Suswillo, R. F., Cheng, M. Z., Tavares, I. A., and Lanyon, L. E. (1997) 
Early responses to dynamic strain change and prostaglandins in bone-derived cells 
in culture. /. Bone Miner. Res. 12, 769-777 '. 

49. Tanaka, S. M. (1999) A new mechanical stimulator for cultured bone cells using 
piezoelectric actuator. /. Biomech. 32, 427-430. 

50. Lanyon, L. E. and Rubin, C. T. (1984) Static vs dynamic loads as an influence on 
bone remodelling. /. Biomech. 17, 897-905. 

51. Rubin, C. T. and Lanyon, L. E. (1985) Regulation of bone mass by mechanical 
strain magnitude. Calcif. Tissue Int. 37, 41 1-417. 

52. Smith, E. L., Martens, F., Roller, K., Clark, W., and Jones, D. B. (2000) The 
effects of 20 days of mechanical loading plus PTH on the E-modulus of cow tra- 
becular bone. /. Bone Miner. Res. 15 (Suppl. 1), S247. 

53. Walker, L. M., Preston, M. R., Magnay, J. L., Thomas, P. B., and El-Haj, A. J. 
(2001) Nicotinic regulation of c-fos and osteopontin expression in human-derived 
osteoblast-like cells and human trabecular bone organ culture. Bone 28, 603-608. 

54. Cheng, M. Z., Zaman, G., Rawlinson, S. C. F., Pitsillides, A. A., Suswillo, R. F., 
and Lanyon, L. E. (1997) Enhancement by sex hormones of the osteoregulatory 
effects of mechanical loading and prostaglandins in explants of rat ulnae. /. Bone 
Miner. Res. 12, 1424-1430. 

55. Cheng, M. Z., Zaman, G., Rawlinson, S. C. F., Suswillo, R. F., and Lanyon, L. E. 
(1996) Mechanical loading and sex hormone interactions in organ cultures of rat 
ulna. /. Bone Miner. Res. 11, 502-51 1. 

56. Rawlinson, S. C. F., Pitsillides, A. A., and Lanyon, L. E. (1996) Involvement of 
different ion channels in osteoblasts' and osteocytes' early responses to mechani- 
cal strain. Bone 19, 609-614. 

57. Rawlinson, S. C. F., Wheeler-Jones, C. P., and Lanyon, L. E. (2000) Arachidonic 
acid for loading induced prostacyclin and prostaglandin E(2) release from osteo- 
blasts and osteocytes is derived from the activities of different forms of phospho- 
lipase A(2). Bone 27, 241-247. 

58. Pitsillides, A. A., Rawlinson, S. C. F., Mosley, J. R., and Lanyon, L. E. (1999) 
Bone's early responses to mechanical loading differ in distinct genetic strains of 
chick: selection for enhanced growth reduces skeletal adaptability. /. Bone Miner. 
Res. 14, 980-987. 

59. Dallas, S. L., Zaman, G., Pead, M. J., and Lanyon, L. E. (1993) Early strain- 
related changes in cultured embryonic chick tibiotarsi parallel those associated 
with adaptive modeling in vivo. /. Bone Miner. Res. 8, 251-259. 

60. Zaman, G., Dallas, S. L., and Lanyon, L. E. (1992) Cultured embryonic bone shafts 
show osteogenic responses to mechanical loading. Calcif. Tissue Int. 51, 132-136. 


62. Jones, D. B. and Scholubbers, J. G. (1987) Evidence that phospholipase C medi- 
ates the mechanical stress effect in bone. Calcif. Tissue Int. 41, 4. 

63. El-Haj, A. J., Minter, S. L., Rawlinson, S. C. F., Suswillo, R. F. L., and Lanyon, 
L. E. (1990) Cellular responses to mechanical loading in vitro. /. Bone Miner. 
Res. 5, 923-932. 

29/Pitsillides/399-422/F1 421 I 2/26/03, 10:56 AM 

I ■ I I I 


Pitsillides et al. 

64. Rawlinson, S. C. F., El-Haj, A. J., Minter, S. L., Tavares, I. A., Bennett, A., and 
Lanyon, L. E. (1991) Loading-related increases in prostaglandin production in 
cores of adult canine cancellous bone in vitro: a role for prostacyclin in adaptive 
bone remodeling? /. Bone Miner. Res. 6, 1345-1351. 

65. Rawlinson, S. C. F., Mohan, S., Baylink, D. J., and Lanyon, L. E. (1993) Exog- 
enous prostacyclin, but not prostaglandin E2, produces similar responses in both 
G6PD activity and RNA production as mechanical loading, and increases IGF-II 
release, in adult cancellous bone in culture. Calcif. Tissue Int. 53, 324-329. 



2/26/03, 10:56 AM 

I ■ I I I 



Extraction of Nucleic Acids from Bone 

Tracy L. Stewart and Val Mann 

1 . Introduction 

Nucleic acids can be extracted from bone for the analysis of gene expres- 
sion, to look for somatic mutations in analysis of tumors or other pathological 
tissue, or for genotyping archive material when other sources of DNA are not 
available. Several methods have been described for the extraction of DNA and 
RNA from tissues (1-4), and several kits based on these methods are currently 
available from biotech companies. If you are extracting DNA from a large 
number samples, however, it is probably cost effective to use a homemade 
method as described here. Successful extraction of nucleic acids from tissue is 
based on four procedures: disrupting the tissue so that extraction reagents can 
reach the cells; disrupting the cell membranes so that nucleic acids are liber- 
ated; separation of the nucleic acid from other cellular components; and pre- 
cipitation and solubilization of the nucleic acid. 

2. Materials (see Note 1) 

2.1. Equipment 

1. Cryogenic mill to pulverize bone samples (Glen Creston, Middlesex, UK). 

2. Sterile scalpels, scissors, and bone cutters. 

3. Sterile 35-mm tissue culture plates. 

2.2. For DNA Extraction 

1. DNA extraction buffer: Add 17.6 mL of 0.75 M sodium citrate, pH 7.0, 26.4 mL 
of 10% sodium lauryl sarkosyl, and 250 g of guanidinium isothiocyanate to 293 
mL of distilled water and mix well. Add 7.2 u,L of |3-mercaptoethanol/mL of lysis 
buffer on the day of use {see Note 2). 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


30/Stewart/423-432/F1 425 1 2/26/03, 10:49 AM 

I ■ I I I 


426 Stewart and Mann 

2. Fumed silica: Required for DNA extraction from dried or embedded bone (see 
Subheading 3.3.)- 

a. Suspend 50 g of unfumed silica (Sigma) in 100 mL of double-distilled water. 
Stir for 1 h and allow to settle under gravity for 1 h. 

b. Aspirate the supernatant and centrifuge at 6000g for 10 min to pellet the glass. 
Resuspend the pellet in 25-30 mL of double-distilled H 2 0. 

c. Add concentrated nitric acid to 50%. Bring to a boil in a fume hood and allow 
to cool. 

d. Pellet the silica by centrifugation at 6000g for 10 min. 

e. Wash the pellet with double-distilled water until the pH of the supernatant is 7.0. 

f. Store the pellet as 50% slurry in double-distilled H 2 (see Note 3). 

3. Silica wash buffer: Add 157.6 mg of Tris-HCl and 70.92 g of guanidinium 
isothiocyanate to 100 mL of distilled water. 

2.3. ForRNA Extraction 

1. Tris-saturated phenol, pH 7.8-8.0 (Sigma). 

2. Hoechst 33258. For DNA quantitation. 

3. RNA extraction buffer: RNAzolB™ (AMS Biotechnology). 

2.4. General Reagents 

1. Diethylpyrocarbonate (DEPC)-treated water: Add DEPC to distilled water to a 
final concentration of 0.1% and incubate at 37°C for 12 h; inactivate the DEPC 
by autoclaving. 

2. Ethidium bromide (10 mg/mL) Dissolve 100 mg ethidium bromide in 10 mL of 
distilled water. Store at room temperature in a bijoux or universal container and 
protect from light by wrapping in aluminum foil. 

3. Chloroform. 

4. Ethanol, 75% and 100%. 

5. Isopropanol, 100%. 

6. 0.5 M ETDA: Add 93.05 g of EDTA to 300 mL of distilled water and add 10 N 
NaOH to pH 8.0. Make up to 500 mL. Autoclave. 

7. 1 M Tris: Add 212 g of Tris base to 800 mL of distilled water. Adjust to pH 8.0 
with concentrated HC1. Make up to 1 L with distilled water. Autoclave. 

8. Tris-EDTA: Add 1 mL of 1 M Tris to 200 \xL of 0.5 M EDTA. Make up to 100 mL 
with distilled water. 

9. 3 M Sodium acetate, pH 5.2: Add 401.8 g of sodium acetate to 800 mL of dis- 
tilled water. Adjust pH to 5.1 with glacial acetic acid. Make up to 1 L with dis- 
tilled water. Autoclave. 

3. Methods 

3. 1. DNA Extraction from Fresh Bone 

The method detailed below enables extraction of genomic DNA fragments 
(up to 80 Kb) from fresh or freshly frozen bone that contains marrow, and the 

30/Stewart/423-432/F1 426 1 2/26/03, 10:49 AM 

I ■ I I I 


Extraction of Nucleic Acids from Bone 427 

majority of DNA will actually be derived from the marrow cells. Since all 
genomic DNA is equal this is only a problem if a somatic mutation in bone 
cells is investigated. 

1. Collect the bone sample in a sterile container containing phosphate-buffered 
saline (PBS) and transport to the laboratory within 1-2 h (see Note 4). 

2. Place the bone tissue in a clean glass Petri dish. Using bone cutters or a strong 
sharp pair of scissors, isolate a piece of bone measuring about 1 cm 3 and transfer 
to a clean 5-mL bijoux container. 

3. Add 1 mL of DNA extraction buffer and homogenize the tissue with the scissors 
until a slurry is obtained. 

4. Transfer 500 uL aliquots of slurry into screw-capped conical-bottomed 1.5-mL 
Eppendorf tubes. 

5. Add one volume of Tris-saturated phenol, followed by one volume of chloroform 
per tube. Mix well by inverting the tubes a few times or by shaking. Do not 
vortex (see Note 5). 

6. Centrifuge the tubes at 10,000g for 20 min to separate the phases. 

7. Transfer the upper layer to a fresh centrifuge tube (taking note of the volume), 
being careful not to disturb the milky layer at the interface. Repeat steps 5-7 if 
the interface is disturbed. 

8. Add one volume of ice-cold isopropanol and 0. 1 volumes of 3 M sodium acetate 
to the supernatant. Mix well and allow to stand for 15 min on ice. 

9. Centrifuge the tubes at 10,000g for 20 min to pellet the DNA (see Note 6). 

10. Aspirate and discard the supernatant, taking care not to disturb the pellet. Wash 
the sample with 1.75 mL of ice-cold ethanol and centrifuge at 10,000g for 5 min. 
Aspirate and discard the supernatant and then repeat the wash. 

11. Dissolve the DNA pellets in 10-50 uL of water or Tris-EDTA buffer (you can 
pool DNA from the same sample at this stage) and quantitate by spectrophotom- 
etry or with Hoechst 33258 (see Note 7). 

12. Store the sample frozen at -20°C or below. 

3.2. DNA Extraction from Cultured Cells 

Extraction of DNA from a cultured cell layer does not require a homog- 
enized step since the cells will be lysed by the salt and detergent in the lysis 

1. Remove the medium from the cultured cells and discard. 

2. Wash the cell layer with sterile PBS, aspirate, and discard. Repeat once. 

3. Add 1 mL of DNA extraction buffer per T-75 flask of cells. Swirl the lysis buffer 
around the flask so that all cells are coated. Leave for 1 min, and detach the cells 
using a cell scraper. 

4. Transfer the extracted cell layer into screw-capped conical-bottomed 1.5-mL 
Eppendorf tubes in 500-fxL aliquots. Pipet the mixture up and down to break up 
the cells. 

5. Follow steps 5-12 of the extraction procedure described in Subheading 3.1. 

30/Stewart/423-432/F1 427 I 2/26/03, 10:49 AM 

I ■ I I I 


428 Stewart and Mann 

3.3. DNA Extraction from Dried or Embedded Bone 

Dried bone, e.g., archeological samples, contain little cellular material and the 
following method is more suitable for these specimens. Be aware that the major- 
ity of DNA obtained will be mitochondrial DNA. The technique relies on the 
propensity of nucleic acids to adsorb to silica (5). Kits based upon this methodol- 
ogy are available, e.g., Silica Adsorption Kit (Roche, East Sussex, UK). 

1 . Homogenize the bone sample with a liquid nitrogen cooled powder mill or manu- 
ally by cooling the specimen with liquid nitrogen and then grinding to a fine 
powder in a mortar and pestle or an electric food grinder. See also Fig. 3 in the 
chapter by Ireland, this volume. 

2. Add 500 u,L of DNA extraction buffer to a clean Eppendorf tube and place on ice. 
Carefully add 100-500 mg of bone powder to the extraction buffer, ensuring that 
it moistens all the powder. 

3. Place the sample on an end-over-end rotator in a cold room for 24 h. 

4. Centrifuge the sample at 10,000g for 5 min in a microcentrifuge to pellet the 
bone powder. Collect the supernatant, containing the DNA, and pipet into a fresh 
Eppendorf tube. 

5. Add 50 [ih of fumed silica to the supernatant. Place on an end-over-end rotator in 
a cold room for 1 h. 

6. Pellet the silica (which contains the adsorbed DNA [5]) by centrifugation at 6000g 
for 5 min and discard the supernatant (see Note 8). 

7. Add 1 mL of silica wash buffer to the pellet and invert a couple of times to resus- 
pend the silica. Centrifuge at 6000g for 5 min and discard the supernatant. 

8. Resuspend the pellet in 95% ethanol (to remove salt contamination) and mix by 
inverting two or three times (see Note 9). Centrifuge at 6000g for 5 min and 
discard the supernatant. 

9. Repeat step 8. 

10. Elute the DNA from the silica by resuspending the pellet in up to 100 |xL of 
either 10 mM Tris HC1, pH 8.0, or distilled deionized water, pH >7.5 see 
Note 10). 

11. Incubate the sample at 56°C for 10 min. Centrifuge at 6000g for 5 min to pellet 
the silica and aspirate the supernatant (which contains the DNA) into a fresh tube 
and store frozen at -70°C or below. 

3.4. RNA Extraction from Fresh Bone 

The method given here uses RNAzolB™ reagent, but the same method can 
be used with TRIzol® reagent, allowing isolation of both RNA and DNA from 
the same sample (see also the chapter by Ireland, this volume). 

1 . Collect the bone sample in a sterile container containing PBS and transport to the 
laboratory within 30 min (see Note 11). 

2. Wearing gloves, place a 1-cm 3 sample of bone tissue into a 35-mm sterile Petri 
dish using sterile forceps (see Note 12). 

30/Stewart/423-432/F1 428 1 2/26/03, 10:49 AM 

I ■ I I I 


Extraction of Nucleic Acids from Bone 429 

3. Add 1.5-1.75 mL of RNAzolB solution to the sample and chop the tissue up into 
a coarse slurry using bone cutters or scissors. Using a sterile scalpel, finely divide 
the sample further until it forms a smooth homogenate. 

4. Carefully transfer up to 1.5 mL/tube of homogenate into a 2-mL screw-cap 
Eppendorf tube and place on ice. 

5. Add 0.1 volume of chloroform per tube and shake vigorously for 15 sec. Place on 
ice for 5 min. 

6. Centrifuge at 10,000g for 20 min at 4°C. 

7. Carefully aspirate the upper layer (taking note of the volume) into a fresh 1.5-mL conical 
bottom screw-top Eppendorf tube, taking care not to disturb the interface (see Note 13). 

8. Add one volume of ice-cold isopropanol and mix gently by inverting the tube 
five or six times. 

9. Place the sample on ice for 15 min to precipitate the RNA (see Note 14). 

10. Centrifuge at 10,000g for 15 min at 4°C to pellet the RNA (see Note 15). Aspi- 
rate and discard the supernatant. 

11. Wash the pellet with 1 mL of ice-cold 75% ethanol and centrifuge at 10,000g for 
5 min at 4°C. Aspirate and discard the supernatant. 

12. Repeat step 11. 

13. Allow the pellet to air-dry (3-5 min). Do not over-dry the pellet. 

14. Dissolve the RNA in 50 uL DEPC-treated distilled water (adjust volume accord- 
ing to sample size). 

15. Quantitate and assess purity of the RNA by spectrophotometry and/or gel elec- 
trophoresis (see Note 16 and Fig. 1). Typical yields of RNA from cells and tis- 
sues are shown in Table 1. 

16. Store the RNA at -70°C until use. 

3.5. Extraction of RNA from Frozen Bone 

If the sample has been stored at -70°C it can be ground to a fine powder 
prior to the extraction protocol, either in a cryogenic mill (Glen Creston, 
Middlesex, UK) or by mortar and pestle (see also Fig. 3 in the chapter by 
Ireland, this volume). 

1 . Pulverize the sample to a fine powder using a cryogenic mill or mortar and pestle 
(see Note 17). 

2. Transfer the powder into a 2-mL Eppendorf tube, add 1.5-1.75 mL of RNAzolB, 
mix well by shaking the tube vigorously, and place on ice. 

3. Proceed through steps 5-16 of the RNA extraction procedure described in Sub- 
heading 3.4. 

3.6. Extraction of RNA from Cultured Cells 

1. Aspirate the medium from the cell culture and discard. 

2. Wash the cells with 10 mL of sterile PBS and discard. 

3. Add 1 .5 mL of RNAzolB per T-75 flask of cells and allow to spread over the cell 
layer. For cells in suspension, see Note 18. 

30/Stewart/423-432/F1 429 I 2/26/03, 10:49 AM 

I ■ I I I 


Stewart and Mann 



Fig. 1. Gel electrophoresis of RNA extracted from bone, bands corresponding to 
the 285S and 185S units of ribosomal RNA should be clearly visible. 

Table 1 

Typical Yields of Total RNA Extracted Using the Described 

Method from Either Tissue or In Vitro Cell Cultures 



260 nm 


280 nm 




1 cm 3 Bone biopsy 
Confluent T75 flask 






Scrape cells using a cell scraper and aspirate the resulting slurry into a 2.0-mL 
Eppendorf tube. 
5. Pipet the slurry up and down through a plOOO to lyse the cells. 

Proceed through steps 5-15 of the RNA extraction procedure described in Sub- 
heading 3.4. 


4. Notes 

1. Safety: The methods described involve use of mercaptoethanol, chloroform, and 
phenol and should be carried out in a fume hood to minimize exposure to these 
solvents. DEPC is a suspected carcinogen and appropriate precautions in han- 



2/26/03, 10:49 AM 

I ■ I I I 


Extraction of Nucleic Acids from Bone 431 

dling it should be taken. We recommend purchase of ethidium bromide in liquid 
to minimize exposure to this mutagen. 

2. All chemicals should be of molecular biology grade. The solutions can be stored 
at 4°C for up to 3 mo. 

3. Working stocks can be stored for up to 1 wk at 4°C. Store stocks frozen at -70°C. 

4. If the DNA extraction is not initiated immediately, freeze the sample at -20°C or 
below for later use. 

5. Vortex-mixing causes long strands of DNA to shear. 

6. Orientate the Eppendorf tube so that you can identify where the DNA pellet lies. 
A pellet should be visible the bottom of the tube. 

7. Dilute the sample 1 :200 for DNA and 1 : 100 for RNA in distilled water and read 
the absorbance at 260 nm and 280 nm using a quartz cuvette in a UV spectropho- 
tometer. Hoechst 33258 is a DNA-specific dye that can be used to quantitate 
DNA. Quantification is achieved by setting up a standard curve of DNA at known 
concentrations (purchase stock with known concentrations) and analyzing the 
test samples in a fluorimeter at an excitation wavelength of 350-363 nm and a 
detection wavelength of 410-480 nm. Another dye useful for high-throughput 
applications is PicoGreen®. The integrity of DNA and RNA can be checked by 
PCR for a housekeeping gene. 

8. You may wish to keep the supernatant in a fresh tube until you are confident that 
the extraction has been successful. 

9. Do not simply add the ethanol and immediately decant it off again; the pellet needs 
to be well mixed so that the ethanol can penetrate the sample and dissolve the salt. 

10. The pH must be above 7.5 to elute the DNA from the silica. 

11. A major challenge in extracting RNA from any tissue bone is to isolate the nucleic 
acid in an intact state before it is degraded. If it is impractical to begin the extrac- 
tion within 30 min, the bone should be snap frozen in liquid nitrogen immedi- 
ately and stored at -70°C or below until use. 

12. All procedures should be carried out wherever possible in an RNase-free envi- 
ronment. Gloves should be worn at all times when handling samples and fre- 
quently changed after handling samples, and all solutions should be prepared to 
ensure that they are free of RNases. This involves making up the solutions for 
RNA use with DEPC-treated water using glassware, stirrers and spatulas that 
have been oven baked for 2 h to inactivate RNases. If possible a separate room or 
area within the laboratory should be dedicated to RNA work and dedicated sets 
of pipets purchased. Surfaces should be routinely cleaned with RNase-inactiva- 
tion reagents such as RNase-Away™ (Fisher Scientific, Leicestershire, UK). The 
use of RNase-free disposable plasticware and filtertips is recommended. Further 
details can be found in Sambrook's text (6). 

1 3 . The upper (clear) aqueous layer contains the RNA and is separated from lower (blue) 
phase containing debris and DNA by a white layer containing protein and lipids. 

14. The samples can be stored overnight at -20°C at this point. Addition of glycogen 
(20 |xg/tube; Boeringer Mannheim) can improve precipitation of small amounts 
of RNA and help in identification of the pellet (see Note 15). 

30/Stewart/423-432/F1 431 1 2/26/03, 10:49 AM 

I ■ I I I 


432 Stewart and Mann 

15. It is a good idea to mark the side of the Eppendorf tube so you can identify where 
the RNA pellet lies. 

16. If the RNA is pure, the optical density reading at 260 nm/280 nm should be 1 .8 or 
greater. If spectrophotometry readings indicate that the RNA is not of the desired 
purity add 1.5 mL of RNAzolB to the RNA solution, mix well, and reextract by 
repeating steps 5-15. We perform such double extractions routinely on bone 
biopsies. DNase treatment can be performed to remove any contaminating 
genomic DNA prior to RT-PCR. We use DNase I amplification grade 
(Invitrogen). Follow manufacturer's instructions. 

17. A small amount of liquid nitrogen can be poured into the pestle to maintain low 

18. For cells in suspension, pellet the cells by centrifugation at 400g for 10 min, 
aspirate the medium, and discard. Suspend the pellet in 1.5 mL of RNAzolB, by 
aspirating up and down using a pi 000 pipet tip and proceed through steps 5-15 
of the extraction described in Subheading 3.4. 


1. Chomczynski, P. and Sacchi, N. (1987) Single step method of RNA isolation by 
acid guanidinium fhiocyanate-phenol-chloroform extraction. Anctlyt. Biochem. 
162, 156-159. 

2. Blin, N. and Stafford, D. W. (1976) A general method for isolation of high 
molecular weight DNA from eukaryotes. Nucl. Acid Res. 3, 2303-2308. 

3. Kupiec, J. J., Giron, M. L., Vilette, D., Jeltsch, J. M., and Emanoil-Ravier, R. 
(1987) Isolation of high molecular weight DNA from eukaryotic cells by 
formamide treatment and dialysis. Analyt. Biochem. 164, 53-59. 

4. Bowtell, D. D. L. (1987) Rapid isolation of eukaryotic DNA. Analyt. Biochem. 
162, 463-465. 

5. Vogelstein, B. and Gillespie, D. (1979) Preparative and analytical purification of 
DNA from agarose. Proc. Natl. Acad. Sci. USA 76, 615-619. 

6. Sambrook, J., Fritsch, E. F., and Maniatis, T. (2000) Molecular Cloning: A Labo- 
ratory Manual, 3rd edit. Cold Spring Harbor Laboratory Press, Cold Spring Har- 
bor, NY. 

30/Stewart/423-432/F1 432 I 2/26/03, 10:49 AM 

I ■ I I I 



Analysis of Gene Expression in Bone 
by Quantitative RT-PCR 

Deborah Ireland 

1 . Introduction 

1.1. RNA Isolation 

The isolation of undegraded RNA, free from inhibitors of reverse transcrip- 
tion-polymerase chain reaction (RT-PCR), is a major technical challenge in 
the analysis of gene expression in the skeleton. Bone is a mineralized tissue 
containing an abundant matrix, which makes RNA isolation difficult. On the 
positive side, however, frozen bone is quite brittle and can be ground to a pow- 
der, thus releasing the cell contents. Reno and colleagues (1) evaluated 12 dif- 
ferent protocols for isolating total RNA from small amounts of rabbit ligament, 
cartilage, and tendon using phenol-guanidinium-based reagents. Like bone, 
these tissues contain few cells in an abundant organic matrix. Absence of ribo- 
somal bands, nondetection of glyceraldehyde-3-phosphate dehydrogenase 
(GAPDH) mRNA on Northern blotting, presence of DNA and/or protein con- 
tamination, and insoluble RNA pellets generated by these procedures led to the 
development of the Trispin method, based on a combination of two commer- 
cially available kits (2). In this procedure, the tissue is powdered in a stainless 
steel ball mill vessel that is cooled in liquid nitrogen. We have used a ball mill 
and modified Trispin method to extract high-quality, RT-PCR-ready RNA from 
bone samples. 

1.2. Quantitation of Gene Expression 

The small quantities of RNA that can be obtained from bone biopsies make 
quantitative RT-PCR an appropriate technique for gene expression studies. 
Several quantification strategies can be used but real-time fluorescence-based 
RT-PCR has the advantage that it allows measurements to be performed over a 

From: Methods in Molecular Medicine, Vol. 80: Bone Research Protocols 
Edited by: M. H. Helfrich and S. H. Ralston © Humana Press Inc., Totowa, NJ 


31/lreland/433-440/F1 433 1 2/26/03, 10:50 AM 

I ■ I I I 



forward primer 
5' * 




reverse primer 


Fig. 1. Generation of reporter fluorescence by the 5' nuclease activity of Taq poly- 
merase on a TaqMan® probe during DNA amplification. The reporter dye is released 
from the annealed probe during extension of the forward primer. Fluorescence is mea- 
sured during each PCR cycle. 

wide range of RNA concentrations with no post-amplification steps. Reaction 
mixtures contain amplification primers and an internal probe, which is labeled 
at the 5' end with a reporter dye and at the 3' end with a quencher molecule that 
suppresses reporter fluorescence in the intact probe. The 3' end of the probe is 
blocked to prevent extension during PCR. The 5' exonuclease activity of Taq 
DNA polymerase cleaves the reporter dye from the annealed probe during 
amplification primer extension, resulting in increased fluorescence (Fig. 1). 
Fluorescence is measured at each PCR cycle and the number of cycles needed 
to reach a threshold level (C t ) determined. Dilutions of standard RNA are used 
to generate a standard curve by plotting C, against log standard RNA concen- 
tration (Fig. 2). Relative or absolute concentrations of sample RNAs can then 
be calculated from the curve. 

2. Materials 

2. 1. RNA Extraction and Solubilization (see Note 1) 

1. Mikro-Dismembrator S ball mill (B. Braun Biotech International, Melsungen, 
Germany) for grinding frozen bone samples. 

2. A 5-mL stainless steel ball mill vessel containing a 10-mm steel/chrome grinding ball. 

3. Stainless steel tray. 

4. Two small dishes suitable for holding liquid nitrogen. 

5. Small stainless steel hammer for crushing bone. 

6. Aluminum foil. 

7. Forceps for handling the bone and bone fragments. 



2/26/03, 10:50 AM 

I ■ I I I 


Gene Expression in Bone 



Rr; vs Cycles 

Standard Curve 


•io -a 

■6 A 


1 3 S 7 9 11 13 16 IT 19 21 23 26 27 29 J1 33 35 17 : 

Cycfe Number 

Fig. 2. TaqMan® plots of log fluorescence against cycle number for 10-fold dilu- 
tions of in vitro transcribed mRNA (A) and threshold cycle number (C t ) against log RNA 
concentration (B). The threshold level is chosen to be in the middle of the linear range. 

8. TRIZOL® (Invitrogen, Paisley, Scotland) or a homemade monophasic reagent 
(PIG-B) prepared according to Weber et al. (2) for solubilizing total RNA from 
powdered bone. 

9. Liquid nitrogen for freezing the bone sample. 

2.2. RNA Purification and Quantitation 

1. Chloroform (Analar grade). 

2. SV Total RNA isolation kit (Promega, Southampton, UK). 

3. NAP-5 columns containing Sephadex G-25 (Amersham Pharmacia Biotech UK 
Ltd., Little Chalfont) for removing RT-PCR inhibitors. 

4. 2 M Sodium acetate solution, pH 4.5. 

5. Ethanol: 95% and 75%. 

6. Spectrophotometer to check optical density readings. 

7. RiboGreen reagent (Molecular Probes, Inc.). 

2.3. Primer Design 

1. Primer Express™ software package (Applied Biosystems, Warrington, UK), for 
primer and probe design. 

2.4. Preparation of In Vitro Transcribed RNA Standards 

1. Thermal cycler for generating PCR products. 

2. Custom oligonucleotide primers. 

3. TA cloning kit (pCR®II TOPO— Invitrogen, Paisley, Scotland) for cloning PCR 
products for in vitro transcription. 



2/26/03, 10:50 AM 

I ■ I I I 


436 Ireland 

4. Wizard Plus SV Minipreps DNA purification system. 

5. SP6/T7 RNA polymerases for making mRNA standards. 

6. Microspin S-300 HR columns (Amersham Pharmacia Biotech UK Ltd., Little 
Chalfont) for purifying in vitro transcribed mRNA. 

7. Molecular Probes RiboGreen™ kit (Cambridge Bioscience, Cambridge, UK) for 
quantifying RNA. 

2.5. Real-Time Fluorescence-Based Quantitative RT-PCR 

1. Gene Amp® 5700 Sequence Detection System (Applied Biosystems, 
Warrington, UK). 

2. Custom oligonucleotide primers (Invitrogen, Paisley, Scotland). 

3. Labeled probes (Applied Biosystems, Warrington, UK). 

4. TaqMan® One-Step RT-PCR Master Mix Reagents Kit (Applied Biosystems, 
Warrington, UK). 

3. Methods 

3. 1. Pulverizing the Bone Sample 

An overview of the procedure is depicted in Fig. 3. 

1. Assemble the apparatus and cool the ball mill vessel and bone sample for several 
min by placing in a dish containing liquid nitrogen (see Note 1). 

2. Wrap the bone sample in foil and hammer until flat, taking care not to pierce the 
foil, and refreeze in the liquid nitrogen (see Note 2). 

3. Quickly and carefully transfer the frozen bone fragments to the Mikro- 
Dismembrator vessel. 

4. Shake for 2 min at maximum speed. 

5. Carefully open the vessel and add 5 mL of TRIZOL® or PIG-B reagent to the 
bone powder (see Note 3). 

6. Transfer the TRIZOL/bone suspension to microcentrifuge tubes and briefly cen- 
trifuge at 12,000g in a centrifuge for 15 min to pellet the insoluble material. 

3.2. RNA Extraction 

1 . Transfer 800 jxl of supernatants to individual microcentrifuge tubes each contain- 
ing 0.25 volumes (200 \iL) of chloroform, and shake vigorously for 15 sec. 

2. Incubate for 3 min, then centrifuge for 15 min at 10,000g to separate the phases. 

3. Transfer the aqueous phase (approx 500 uL) to microcentrifuge tubes containing 
0.4 volumes (200 uL) of 95% ethanol and mix by pipetting three to four times. 

4. Transfer the contents of the microcentrifuge tubes to spin columns and continue 
with RNA purification according to the manufacturer's instructions, using 100 uL 
of nuclease-free water to elute RNA from each spin column. 

5. Combine the eluates from each spin column in a single microcentrifuge tube. 

6. Add the RNA solution to an NAP-5 column preequilibrated with 10 mL of nu- 
clease-free water. When all solution has entered the gel bed, elute the RNA with 
1 mL of nuclease-free water, collecting 100 \xL fractions. 

31/lreland/433-440/F1 436 1 2/26/03, 10:50 AM 

I ■ I I I 


Gene Expression in Bone 


ball mill 




containing liquid 



foil packet 

n _&p>*mh 





I (6 
N N N, "V ' 



Fig. 3. Overview of RNA extraction from bone. 

7. Determine RNA concentration and purity of fractions by measuring absorbance 
at 260/280 nm (see Note 4). Integrity of the RNA can also be checked by gel 
electrophoresis (Fig. 4). 

8. Combine fractions containing RNA and precipitate using 1/10 volume of 3 M 
sodium acetate solution, pH 4.5, and 2.5 volumes 95% ethanol. Leave on ice for 
10 min or at -20°C overnight. 

9. Centrifuge at 10,000-12,000g for 15 min to pellet RNA. 

10. Carefully wash pellet with 75% ethanol and air-dry. 

11. Resuspend RNA in nuclease-free water and store at -80°C. 



2/26/03, 10:50 AM 

I ■ I I I 





4 5 6 7 

fraction number 

Fig. 4. (A) Yield and OD 2 6o/280 ratios of total RNA purified from 500 mg of frozen 
iliac crest biopsy using the Tripsin method described. The RNA was eluted as 100-jxL 
fractions using 1 mL of nuclease-free water. The integrity of the RNA is apparent on 
electrophoresis of an aliquot through a 1.2% formaldehyde-agarose gel. 

3.3. Primer Selection for Cloning and RT-PCR 

1. Fetch the cDNA sequence of interest from the Embl or Genbank data bases. 

2. Obtain the corresponding gene sequence and mark the exon-exon boundaries 
onto a copy of the cDNA sequence. 

3. Export the cDNA sequence onto the sequence page in the Primer Express file 
TaqMan™ probe and primer design. 

4. Mark the exon-exon boundaries using the program junction tool. 

5. Using the probe tool, select a probe sequence that is centred on an exon-exon 
boundary and has a T m of 70°C. Follow the manufacturer's instructions and make 
sure that the probe has more Cs than Gs and that it does not have a G at the 5' end. 
The probe can be complementary to either strand of the PCR product. 

6. Using the default parameters, allow the program to select suitable amplification 
primers. Disregard any that have fewer than three A/Ts among the five bases at 
the 3' end. 

7. If a satisfactory result is not obtained in step 5, repeat steps 4 and 5 with a differ- 
ent exon-exon boundary. 

8. Design cloning primers that will give a 300-500-basepair PCR product that 
includes the TaqMan amplicon. A suitable restriction enzyme recognition site 
can be added to one of the primers to allow insert orientation if a TA cloning 
vector is used. 

3.4. Generating an Internal Control RNA Molecule 

1. Generate the template for in vitro transcription of the target RNA by PCR. 

2. Clone the product into the pCRII-TOPO vector and check orientation of insert 
(see Note 5). 



2/26/03, 10:50 AM 

I ■ I I I 


Gene Expression in Bone 439 

3. Linearize the plasmid. 

4. Transcribe using RNA polymerase. 

5. Treat the transcription reaction with DNase and remove any free nucleotides on a 
spin column. 

6. Quantify RNA using RiboGreen™ reagent. 

3.5. Real-Time Fluorescence-Based Quantitative RT-PCR 

1 . Create a new file with the desired plate layout and instrument control using the 
Gene Amp® 5700 Sequence Detection System software. Relative or absolute 
quantities must be assigned to the standards (see Note 6). 

2. Prepare dilutions of the appropriate in vitro transcribed mRNA in nuclease-free 

3. Prepare enough diluted TaqMan® One-Step RT-PCR master mix containing 
probe and primers for the desired number of replicates of both control and sample 
reactions (see Note 7). 

4. Add control and sample RNA to the diluted master mix and plate out into a 96- 
well PCR plate according to the chosen layout. 

5. Cap the wells of the plate with optical caps, insert the plate into the thermal 
cycler, and run the assay. 

6. When the run is completed, look at the amplification plot showing fluorescence 
signals against cycle number and set a baseline. Change the ^-axis to a log scale 
and choose a threshold value in the middle of the linear range. 

7. Edit the standards to include only those dilutions that give a linear standard curve 
(see Note 8). 

8. Use the program to generate a report of the assay. 

9. Assay the same samples for an internal control RNA (see Note 9). 

10. Normalize the amounts of specific mRNA in the samples by dividing by the 
amounts of internal control RNA. 

11. Calculate the standard deviations of the quotients according to the formula 

cv =VCV 1 2 + CV 2 2 

where CV = standard deviation ( mean value. 

12. Express amounts of mRNA in the samples as A-fold differences from the amount 
in the experimental control or calibrator (3). 

13. Assess the significance of the differences between sample means using the 
approximate test for unequal variances based on the t distribution (4). 

4. Notes 

1. Preparation of apparatus: It is important that all the apparatus is wrapped in alu- 
minum foil before use and heated at 180°C for 8 h to sterilize it and to destroy 

2. Size of bone sample: The bone sample should be <500 mg. 

3. Pulverizing the tissue: Make sure that the bone is completely powdered — if not, 
carefully replace the vessel lid, cool the vessel and contents in liquid nitrogen, 

31/lreland/433-440/F1 439 1 2/26/03, 10:50 AM 

I ■ I I I 


440 Ireland 

remount them on the Mikro-Dismembrator, and shake at maximum speed for 
2 min. Note that when the TRIZOL® or PIG-B reagent is added to the vessel, the 
reagent may freeze. If this happens it should be allowed to thaw at room tempera- 
ture and then incubated for a further 5 min at room temperature to ensure that 
nucleoprotein complexes are dissociated. 

4. Spectrophotometry of fractions from NAP columns: The ratio of absorbance at 
260/280 nm of fractions containing purified RNA should be 1.8-2.0. 

5. Cloning products for internal mRNA control: We use the pCRII-TOPO vector, 
but many alternative cloning vectors and other PCR-based strategies can be used 
to generate in vitro transcribed mRNA. Orientation of the insert can be checked 
by DNA sequencing or restriction enzyme analysis. 

6. Setting up standard curve on Gene Amp 5700: Quantities must be assigned to the 
standards so that a standard curve can be generated. 

7. Avoiding PCR contamination: Remember to include no-template controls to 
check for contamination. It is important to avoid contamination in all PCR 
experiments: we recommend using a laminar flow cabinet for the preparation of 
the reactions. The number of identical and experimental replicates must be 
decided according to the expected differences in mRNA levels. Identical repli- 
cates will give an idea of the precision of the assay. 

8. Screening for PCR inhibitors: Because RNA isolated from bone samples can con- 
tain RT-PCR inhibitors, it is important to prepare dilutions of such samples to 
check for inhibition. 

9. Choice of internal control RNA: The internal control is usually a housekeeping 
gene such as glyceraldehyde-3-phosphate dehydrogenase or |3-actin or a riboso- 
mal RNA. 


1. Reno, C, Marchuk, L., Sciore, P., Frank, C. B., and Hart, D. A. 1997. Rapid 
isolation of total RNA from small samples of hypocellular, dense connective tis- 
sues. BioTechniques 22, 1082-1086. 

2. Weber, K., Bolander, M. E„ and Sarkar, G. (1998) PIG-B: a homemade monopha- 
sic cocktail for the extraction of RNA. Mol. Biotechnol. 6, 13-11 . 

3. Perkin-Elmer Corporation (1997) User Bulletin No. 2. ABI PRISM 7700 Sequence 
Detection System. Relative Quantitation of Gene Expression. 

4. Armitage, P. and Berry, G. (1994) Statistical Methods in Medical Research. 
Blackwell Science, Boston, pp. 111-113. 

31/lreland/433-440/F1 440 I 2/26/03, 10:50 AM